Skip to main content

MCP tools

The Corte MCP server gives agents the same primitives as the REST API, shaped as tools. It speaks Streamable HTTP at https://api.corte.so/mcp and authenticates either with browser consent (OAuth 2.1) or with a normal API key.

For setup and the conceptual tour, see Connect agents with MCP. This page is the reference: every tool, every parameter, every return shape.

Transport​

POST https://api.corte.so/mcp
Authorization: Bearer corte_at_… (or corte_sk_…)
Content-Type: application/json

JSON-RPC over a single POST. The server is stateless — every request is self-contained, there's no session to establish or resume. GET /mcp and DELETE /mcp return 405.

Connect from Claude Code
claude mcp add --transport http corte https://api.corte.so/mcp

A browser window opens for you to approve what the agent may reach. Any client that follows the MCP authorization spec connects this way; a client that only sends headers uses an API key instead.

Connect with OAuth​

/mcp is an OAuth 2.1 protected resource. An unauthenticated call answers 401 with

WWW-Authenticate: Bearer resource_metadata="https://api.corte.so/.well-known/oauth-protected-resource"

which is enough for a client to find everything else on its own:

EndpointWhat it is
/.well-known/oauth-protected-resourceNames the resource and its authorization server (RFC 9728)
/.well-known/oauth-authorization-serverEndpoints, grants and scopes (RFC 8414)
POST /oauth/registerDynamic client registration (RFC 7591) — public clients, token_endpoint_auth_method: none
GET /oauth/authorizeSign-in and the consent screen
POST /oauth/tokenCode exchange and refresh

PKCE with S256 is required. Redirect URIs must be https, or http on loopback (localhost, 127.0.0.1, ::1) on any port. Access tokens (corte_at_…) last an hour; the refresh token rotates on every use and lasts 30 days.

A client that names no scope is granted read-write across every area except your account, which stays read-only. account:rw is never issued to a connected app — an app cannot mint API keys, change your plan, or disconnect its neighbours.

What's different from REST​

Nearly every tool wraps a REST primitive, with three deliberate differences:

Tools wait by default. generate, run_application, and run_workflow block on completion up to a waitSeconds budget, then return whatever state they reached. Agents work better with one call than with a polling loop they have to invent. Pass waitSeconds: 0 to return immediately.

Results are JSON text. Every tool returns a single text content block containing pretty-printed JSON. Parse it as JSON.

Most errors are plain messages. Tool failures surface as MCP errors carrying the REST error message — Workflow not found, Unknown model, and so on. The structured { error: { code } } envelope isn't preserved. The exception is anything the account itself blocked: an empty balance, an unconfirmed address, a paid-only model, full storage. Those come back as a readable result — see When the account blocks the call.

:::warning Agents spend real credits MCP tools run as your account. Every generation or workflow run an agent starts costs exactly what clicking Generate would. Give agents their own key so you can revoke them independently, and watch Settings → Usage. :::


Account​

get_credits​

Current credit balance. Cheap; an agent should call it before expensive work.

Input — none.

Returns

{
"monthlyBudgetCredits": 3500,
"purchasedCredits": 500,
"spentCreditsThisPeriod": 1288,
"remainingCredits": 2712
}

list_projects​

The projects the account can work in — its own and the shared ones it is a member of. This is how an agent turns "in my Playground project" into the projectId every other tool accepts; without one, work lands in and bills to defaultProjectId. Needs account:read.

No parameters.

Returns { projects, defaultProjectId }. Each project carries id, name, description, role (owner or contributor), owner ({ email, name }), remainingCredits (what the project can spend — a shared project spends its owner's credits) and updatedAt.


get_account​

The authenticated account: plan, wallet, storage, and the gates that decide whether a spend goes through. For the balance alone, get_credits is cheaper and takes a projectId. Needs account:read.

Input — none.

Returns

{
"email": "sam@example.com",
"name": "Sam",
"plan": {
"tier": "pro",
"name": "Pro",
"monthlyPriceUsd": 30,
"annualPriceUsd": 288,
"monthlyBudgetCredits": 3500,
"billingInterval": "month",
"currentPeriodEnd": "2026-10-14T09:12:00.000Z",
"renewsAt": "2026-10-14T09:12:00.000Z",
"cancelAtPeriodEnd": false,
"pastDue": false
},
"credits": {
"remaining": 2712,
"monthlyBudget": 3500,
"purchased": 500,
"spentThisPeriod": 1288,
"firstTopupBonus": null
},
"storage": { "usedBytes": 41943040, "quotaBytes": 536870912000, "usedPercent": 0 },
"emailVerified": true,
"hasPaid": true,
"suspended": false,
"canSpend": true,
"billingEnabled": true,
"settingsUrl": "https://app.corte.so/settings"
}

plan.name is what the plan pickers call the tier (max is Business). credits.firstTopupBonus is the one-time bonus a first top-up still carries, or null once any purchase has been made. hasPaid is true after any top-up or plan — the same purchase that unlocks the models list_models marks paidOnly. canSpend is the gate every spend passes: a suspended account never spends; an unconfirmed address spends only once it has paid. Storage is a hard quota — at the cap, uploads and imports block until files are deleted or the plan grows; generation results always save. Shared projects bill their owner, so this describes the caller's own account, never a project's payer.


get_usage​

What the account spent, by kind and by day, over a window — plus a page of the activity feed. Needs account:read.

ParameterTypeRequiredDescription
daysintegernoWindow ending now, 1–366. Default 30. Ignored when from is set.
fromstringnoWindow start, ISO-8601. Overrides days.
tostringnoWindow end, ISO-8601, exclusive. Defaults to now.
limitintegernoActivity lines to include, 1–200. Default 20.
offsetintegernoActivity lines to skip.

Returns

{
"from": "2026-08-27T09:00:00.000Z",
"to": "2026-09-26T09:00:00.000Z",
"spentCredits": 1288,
"grantedCredits": 3500,
"byKind": [
{ "kind": "generation", "spentCredits": 940, "grantedCredits": 0, "entries": 61 },
{ "kind": "build_agent", "spentCredits": 302, "grantedCredits": 0, "entries": 188 },
{ "kind": "transcription", "spentCredits": 46, "grantedCredits": 0, "entries": 4 },
{ "kind": "monthly_grant", "spentCredits": 0, "grantedCredits": 3500, "entries": 1 }
],
"daily": [
{ "day": "2026-09-25", "generations": 120, "agents": 40, "compute": 6, "other": 0 }
],
"recent": [
{
"id": "…", "kind": "build_agent", "credits": 38, "note": "Build: Shop landing", "ref": "…",
"at": "2026-09-25T17:02:11.000Z", "charges": 42, "firstAt": "2026-09-25T15:40:03.000Z",
"projectId": "…", "context": { "surface": "build", "appId": "…", "appName": "Shop landing" }
}
],
"recentTotal": 214,
"remainingCredits": 2712
}

byKind is one row per ledger kind that moved in the window, biggest spend first. daily buckets each UTC day's debits into generations, agents (copilot chats, Build turns, agent runs), compute (sandbox time) and other; days with no spend are absent. recent is the same feed the usage page shows — one Build session or agent run is one line however many charges it settled (charges and firstAt appear when it folded more than one) — and it pages over the whole history, not just the window. credits on a line is positive for a spend and negative for a grant. The feed's thumbnails are not included. Shared-project spend lands on whoever pays for the project, so this is what this account paid, including its collaborators' work; one project's own breakdown is GET /v1/projects/:id/usage.


list_plans​

The plan ladder, cheapest first, with the account's current plan marked — so an agent can say what an upgrade costs and what it adds. Needs account:read.

Input — none.

Returns

{
"currentTier": "none",
"plans": [
{ "tier": "none", "name": "Free", "monthlyPriceUsd": 0, "annualPriceUsd": null, "annualMonthlyPriceUsd": null, "monthlyBudgetCredits": 0, "storageQuotaBytes": 5368709120, "current": true },
{ "tier": "standard", "name": "Standard", "monthlyPriceUsd": 10, "annualPriceUsd": 96, "annualMonthlyPriceUsd": 8, "monthlyBudgetCredits": 1000, "storageQuotaBytes": 107374182400, "current": false },
{ "tier": "pro", "name": "Pro", "monthlyPriceUsd": 30, "annualPriceUsd": 288, "annualMonthlyPriceUsd": 24, "monthlyBudgetCredits": 3500, "storageQuotaBytes": 536870912000, "current": false },
{ "tier": "max", "name": "Business", "monthlyPriceUsd": 80, "annualPriceUsd": 768, "annualMonthlyPriceUsd": 64, "monthlyBudgetCredits": 10000, "storageQuotaBytes": 1099511627776, "current": false }
],
"billingEnabled": true
}

tier is the id create_subscription_checkout takes. Yearly bills once at 20% off and credits still arrive monthly; annualMonthlyPriceUsd is the per-month figure the plan pickers show. Any plan — like any top-up — unlocks the models list_models marks paidOnly; editing and export are free on every plan. Purchased credits never expire; a plan's monthly credits reset each period.


Credits​

An agent that runs out of credits mid-task can hand you the way to fix it. All three tools return a link; you open it in your browser and pay, or change your billing, there. The card never goes near the agent, and no tool moves money on its own — until somebody pays the link or acts in the portal, nothing has happened.

All three need account:read. Creating a link neither reads your account nor changes it, so it doesn't cost the account:rw grant that also mints API keys.

create_topup_checkout​

A payment link for a one-off top-up. 100 credits per dollar; purchased credits never expire.

ParameterTypeRequiredDescription
dollarsintegeryesWhole dollars, 5–1,000. Buys dollars × 100 credits.

Returns

{
"url": "https://…",
"dollars": 25,
"credits": 2500,
"firstTopupBonusCredits": 250,
"expiresAt": "2026-08-31T14:00:00.000Z"
}

firstTopupBonusCredits is the one-time bonus this account still has coming on its first purchase, or 0 once it's been collected. expiresAt is when the link stops working, or null where the provider doesn't say.

Credits land within about a minute of payment. Poll get_credits rather than assuming, and don't retry the work that failed until the balance has actually moved.

create_subscription_checkout​

A payment link for a plan. Yearly bills once at 20% off; credits still arrive monthly.

ParameterTypeRequiredDescription
tier"standard" | "pro" | "max"yesWhich plan to buy.
interval"month" | "year"noDefaults to month.

Returns { "url": string, "tier": string, "interval": string }.

Asking for the plan the account is already on is refused, the same way POST /v1/billing/checkout refuses it.

A link to the billing portal, to hand to the user: where they change the card on file, download invoices, switch between monthly and yearly, or cancel. Nothing changes until they act in their browser, so it needs only account:read, like the checkout tools.

Input — none.

Returns { "url": string }.

Refused with forbidden where billing is turned off for the workspace, the same as POST /v1/billing/portal. Use it for "update my card", "send me my invoices", "cancel my plan", or when get_account reports pastDue; to add credits or start a plan, the checkout tools are the direct route.

When the account blocks the call​

Eight failures aren't bugs or bad arguments — the call was fine and the account is what stopped it. Any tool can return one, and it comes back as an isError result whose text is JSON:

{
"error": "insufficient_credits",
"message": "Insufficient credits",
"balance": 12,
"estimatedCredits": 40,
"nextStep": "Ask the user to add credits: call create_topup_checkout({dollars}) and give them the link — they pay in their browser and the credits land within a minute. Then poll get_credits and retry. Do not retry before the balance changes."
}
FieldWhen
errorAlways — the error code.
messageAlways — the same sentence REST would return.
balanceinsufficient_credits and plan_required: the caller's own remainingCredits. Omitted where the payer is somebody else.
estimatedCreditsWhere the refused work knew its cost — a generation's estimate, a transcription's price.
nextStepAlways — one sentence naming what to do, written for the agent.

The eight codes and what each one actually needs:

CodeWhat clears it
insufficient_creditsA top-up or a plan.
owner_insufficient_creditsThe project owner adding credits — this work is billed to them, so buying on your own account won't unblock it.
plan_requiredAny purchase. Pro models unlock on a first top-up or a plan, not on a particular tier.
email_unverifiedClicking the link in the confirmation email. Nothing to buy.
owner_email_unverifiedThe project owner confirming theirs.
account_suspended · owner_account_suspendedContacting support.
storage_limitDeleting files, or a plan with more storage. Storage is a hard quota with no overage billing.

nextStep adapts to what the calling key can actually do: a key without account:read is pointed at the app rather than at a checkout tool it can't call, and a workspace with billing turned off is never told to buy anything.


Projects​

A project is where media, cuts, workflows, generations and content sit together, with its own members and its own spend. list_projects (above) is how an agent finds one; the tools here read and manage a single project. Every tool takes an optional projectId; without one it acts in the project the credential works in (the account's default project when none is pinned). A project the caller is not a member of reads as Unknown project — its existence is not confirmed. Need projects:read for the reads and projects:rw for the rest.

Deleting a project, handing it to another member and moving it between organisations have no tools: they change who owns work, or destroy it, and stay in the app.

get_project​

One project in full: the project, every member, and — for the owner — the invites still open.

ParameterTypeRequiredDescription
projectIdstringnoFrom list_projects. Omit for the credential's project.

Returns { "project": Project, "members": Member[], "invites"?: Invite[], "url": string }. project carries role (owner or contributor), paidBy ({ kind: "owner" | "member" | "org", name, email }) and remainingCredits — what the project can spend right now. Each member is { userId, email, name, role, joinedAt }; userId is what remove_project_member takes. invites is present only for the owner. url is the app's project page.

Returns
{
"project": {
"id": "8e1c4a52-2b0f-4c7e-9d3a-6f2b1c8e4d70",
"name": "Spring campaign",
"description": null,
"avatarUrl": null,
"role": "owner",
"orgId": "org_personal_…",
"owner": { "email": "you@example.com", "name": "You" },
"paidBy": { "kind": "owner", "name": "You", "email": "you@example.com" },
"remainingCredits": 2712,
"createdAt": "2026-09-26T14:10:02.114Z",
"updatedAt": "2026-09-26T14:10:02.114Z"
},
"members": [
{ "userId": "u_1", "email": "you@example.com", "name": "You", "role": "owner", "joinedAt": "2026-09-26T14:10:02.120Z" }
],
"invites": [],
"url": "https://app.corte.so/settings/project"
}

create_project​

Make a new project, owned by the caller.

ParameterTypeRequiredDescription
namestringyes1–200 characters.
orgIdstringnoA team organisation to create it in; omit for the caller's personal space.

Returns the same shape as get_project: the project (its id is what every other tool's projectId takes), members (the caller, as owner), invites: [] and url.

The caller pays for work done in the project unless it lives in a team whose wallet pays. Agents should not create projects the person did not ask for — most work belongs in the one they already have.

update_project​

Rename a project, change its description, or set its picture. Owner only.

ParameterTypeRequiredDescription
projectIdstringnoOmit for the credential's project.
namestringnoA new name, 1–200 characters.
descriptionstring | nullnoUp to 2000 characters. Empty or null clears it.
avatarStorageIdstring | nullnoAn image the caller can read (a storageId); null removes the picture.

Omitted fields are left alone. Returns the same shape as get_project. A contributor is refused (forbidden).

invite_to_project​

Invite someone to a project by email. Sends an email from Corte on the caller's behalf. Owner only.

ParameterTypeRequiredDescription
projectIdstringnoOmit for the credential's project.
emailstringyesThe address to invite. Lower-cased; must be a valid email.

Returns { "invite": { id, projectId, projectName, email, invitedBy: { email, name }, createdAt } }.

The invitee joins as a contributor once they accept, with access to everything in the project and their work billed to it. They need no account yet: the invite waits for a signup with that address. Refused when the address is already a member or already invited, when it is the caller's own, and when the project is full (20 people, members and pending invites together). An agent should confirm the address with the person before calling — a wrong address is an email to a stranger.

cancel_project_invite​

Withdraw an invite that has not been accepted. Owner only.

ParameterTypeRequiredDescription
projectIdstringnoOmit for the credential's project.
inviteIdstringyesFrom get_project's invites, or the invite invite_to_project returned.

Returns { "ok": true, "inviteId" }. Someone who already accepted is a member — use remove_project_member.

remove_project_member​

Take a member out of a project. They lose access at once; what they made stays.

ParameterTypeRequiredDescription
projectIdstringnoOmit for the credential's project.
userIdstringyesThe member's userId from get_project.

Returns { "ok": true, "userId" }. The owner can remove any contributor; a contributor can remove only themselves (leave). The owner cannot be removed (forbidden).

get_project_usage​

What a project has spent: its credit ledger, newest first, with per-member totals over the whole history.

ParameterTypeRequiredDescription
projectIdstringnoOmit for the credential's project.
limitnumbernoEntries per page. Default 50, max 200.
offsetnumbernoEntries to skip. Default 0.

Returns { "entries": LedgerEntry[], "totalsByActor": [{ actorUserId, actorEmail, spentCredits }], "total", "limit", "offset" }. Each entry carries kind, deltaCredits (negative for spend), ref, note, actorUserId, actorEmail and createdAt. totalsByActor covers the project's whole history regardless of the page; total is the number of entries visible to the caller. The owner (and an organisation's billing holder) sees everyone's spend; a contributor sees only their own. For what the project can still spend, get_project has remainingCredits.

Returns
{
"entries": [
{ "id": "le_9", "kind": "generation", "deltaCredits": -3, "ref": "gen-2", "note": null,
"createdAt": "2026-09-26T14:12:40.001Z", "projectId": "8e1c…", "actorUserId": "u_1", "actorEmail": "you@example.com" }
],
"totalsByActor": [{ "actorUserId": "u_1", "actorEmail": "you@example.com", "spentCredits": 10 }],
"total": 2,
"limit": 1,
"offset": 0
}

Models & generation​

list_models​

The generation catalog with capabilities and pricing.

ParameterTypeRequiredDescription
kind"image" | "video" | "audio" | "upscale"noFilter by model kind.

Returns { "models": [...], "total": number } — the same model objects as GET /v1/models, always the full catalog.

generate​

Submit a generation on any model, and wait for it.

ParameterTypeRequiredDescription
modelstringyesModel id from list_models.
paramsobjectyesMust match the model's kind. See Generation parameters.
projectIdstringnoBill to a shared project — the owner pays.
waitSecondsnumbernoDefault 60, max 300. 0 returns immediately.

Returns { "job": GenerationJob, "estimatedCredits": number }.

If the job is still running when the wait budget expires, the job is returned in its current state — poll with get_generation. The wait timing out is not an error.

Example call
{
"model": "nano-banana-2-1",
"params": {
"kind": "image",
"prompt": "Editorial flat lay of a trail-running shoe on slate, cold morning light",
"aspectRatio": "4:5",
"numImages": 2
},
"waitSeconds": 120
}
Returns
{
"job": {
"id": "a91c2d7e-5f80-4c39-b6a2-13e7d0f4c885",
"model": "nano-banana-2-1",
"kind": "image",
"status": "succeeded",
"resultUrls": ["https://media.corte.so/…?X-Amz-Signature=…"],
"resultStorageIds": ["c74a1e08-9b6f-4d13-a205-e8f30b7c9d61"],
"costCredits": 24,
"completedAt": "2026-07-28T21:02:55.410Z"
},
"estimatedCredits": 24
}

get_generation​

Fetch a generation job by id — the polling tool.

ParameterTypeRequiredDescription
jobIdstringyesJob id. Must belong to the key's account.

Returns { "job": GenerationJob }. A finished job carries results[] in place of raw result URLs: each result has its storageId, a downloadUrl on Corte's own host that opens for 24 hours without signing in, and a libraryUrl that is permanent and opens the asset in Corte for the signed-in person — the link to show them. For a permanent public URL, publish_asset.


Transcription​

transcribe_asset​

Turn the speech in a stored audio or video file into subtitles, and wait for them.

ParameterTypeRequiredDescription
storageIdstringyesThe audio or video file — from list_assets, get_asset, upload_asset or a generation's results[].
languagestringnoBCP-47 code of the speech (en, pt-BR). Omit to detect it.
projectIdstringnoBill to a shared project — the owner pays.
durationSecondsnumbernoOnly when the tool asks for it: the length is read from the file, and a file whose header will not give it up needs the caller to say.
waitSecondsnumbernoDefault 120, max 300. 0 returns immediately.

Returns { "job": TranscriptionJob & { durationSeconds }, "text", "language", "segments", "srt", "vtt" }. The transcript fields are present once the job has succeeded: text is the whole transcript, segments the timed, speaker-labelled chunks ({ text, start, end, speaker? }), and srt / vtt the finished subtitle files as strings — save them as .srt / .vtt and they drop straight into any player or editor. Word-level timings are on GET /v1/transcriptions/:id/result.

Cost: 72 credits per hour of media, charged only once the transcript exists. A file that is not audio or video is refused before anything is billed.

If the job is still running when the wait budget expires, the job is returned in its current state — poll with get_transcription.

Example call
{
"storageId": "c74a1e08-9b6f-4d13-a205-e8f30b7c9d61",
"waitSeconds": 180
}
Returns
{
"job": {
"id": "5d1e8c2a-0f7b-4a6e-9c3d-2b8f1e4a7c90",
"status": "succeeded",
"language": "en",
"errorMessage": null,
"costCredits": 4,
"createdAt": "2026-09-26T18:04:11.902Z",
"durationSeconds": 183.4
},
"text": "Welcome back. Today we're looking at…",
"language": "en",
"segments": [
{ "text": "Welcome back.", "start": 0.31, "end": 1.02, "speaker": "speaker_0" },
{ "text": "Today we're looking at…", "start": 1.4, "end": 3.87, "speaker": "speaker_0" }
],
"srt": "1\n00:00:00,310 --> 00:00:01,020\nWelcome back.\n\n2\n…",
"vtt": "WEBVTT\n\n00:00:00.310 --> 00:00:01.020\nWelcome back.\n\n…"
}

get_transcription​

Fetch a transcription job by id — the polling tool.

ParameterTypeRequiredDescription
jobIdstringyesJob id. Visible to the account that started it and to members of its project.

Returns the same shape as transcribe_asset: the job, and once it has succeeded the text, segments, srt and vtt.


Assets​

list_assets​

A project's media — images, video and audio the account uploaded, imported or generated. Reading assets needs only media:read; importing new ones is media:rw.

ParameterTypeRequiredDescription
projectIdstringnoDefaults to the default project.
kind"image" | "video" | "audio"noFilter on the content type's major part.
limit · offsetnumbernoPaging. Default limit 50.

Returns { objects, total }. Each object carries storageId, filename, contentType, sizeBytes, createdAt, origin, a downloadUrl that opens for 24 hours and a permanent libraryUrl. total counts the filtered set, so paging within a kind is safe.

Requires membership of the project — the same check the media picker makes.

get_asset​

One stored file by id: its metadata plus a downloadUrl that opens for 24 hours without signing in, a thumbDownloadUrl for images and video, and a permanent libraryUrl that opens it in Corte.

ParameterTypeRequiredDescription
storageIdstringyesFrom list_assets, upload_asset, or a generation result.

Use it to check what an id actually refers to before placing it in a content document or handing it to a model.

:::note Unknown, not forbidden A file the account neither owns nor shares through a project reads as unknown. That is deliberate: a "forbidden" answer would confirm the id exists, turning the tool into a way to probe for other people's files. :::

upload_asset​

Import media from an https URL into storage. Returns the storageId every other tool accepts as an input.

ParameterTypeRequiredDescription
urlstringyeshttps URL of an image, video, or audio file.
filenamestringnoOverrides the name derived from the URL.

Returns

{
"storageId": "b0e7c9a1-4d52-4f0e-8c31-77a2e5d9f6b8",
"url": "https://media.corte.so/…?X-Amz-Signature=…",
"sizeBytes": 482113
}

There is no MCP tool for uploading local bytes — agents work from URLs. For local files use the REST ticket flow.


Applications​

list_applications​

The curated Applications catalog with each entry's input requirements.

Input — none.

Returns { "applications": [...] } — the same key as GET /v1/applications, with the objects trimmed to what an agent needs to choose and run one:

{
"applications": [
{
"id": "flatlay-to-model",
"name": "Flatlay to Model",
"kind": "image",
"description": "Upload a flatlay of clothing or accessories and get a styled photo of a model wearing the items…",
"inputKind": "image",
"minImages": 1,
"maxImages": 1,
"acceptsText": true,
"presets": [
{ "id": "studio-white", "name": "Studio White" },
{ "id": "street-style", "name": "Street Style" }
]
}
],
"total": 43
}

Preset prompts are omitted — only id and name. Use GET /v1/applications if you need the full objects.

run_application​

Run an Application and wait for the generation.

ParameterTypeRequiredDescription
applicationIdstringyesId from list_applications.
imageStorageIdsstring[]cond.For image-input Applications.
videoStorageIdstringcond.For video-input Applications.
audioStorageIdstringcond.For audio-input Applications (voice cleanup).
durationSecondsnumbercond.Input media length — required for duration-billed Applications (video upscale, voice cleanup).
textstringnoExtra guidance. Only when acceptsText is true.
presetIdstringnoA preset id from the Application.
aspectRatiostringnoOutput shape from the Application model's aspect ratios — "9:16" turns a landscape clip vertical with Reframe Video. Omitted: the model's default.
projectIdstringnoBill to a shared project — the owner pays.
waitSecondsnumbernoDefault 60, max 300. 0 returns immediately.

Returns { "job": GenerationJob, "estimatedCredits": number }.

Validation is identical to POST /v1/applications/:id/run — the two share one implementation.


Content​

Tools over the Content management API — enough for an agent to model a CMS, fill it, publish it, and hand the app a read token. Every tool calls the same code the REST routes do, so permissions, validation and versioning behave identically.

projectId is optional throughout; omitted, it means the account's default project.

list_content_types​

The project's schema registry, each type with its fields and documentCount.

ParameterTypeRequiredDescription
projectIdstringnoDefaults to the default project.
limit · offsetnumbernoPaging.

get_content_type​

ParameterTypeRequiredDescription
contentTypeIdstringyesType id, or its machine name (post).
projectIdstringnoWhere a machine name is looked up. Defaults to the default project.
versionnumbernoA historical schema version.

create_content_type​

ParameterTypeRequiredDescription
projectIdstringnoDefaults to the default project.
namestringyescamelCase machine name. Immutable afterwards.
titlestringyesWhat editors see.
descriptionstringnoFree text.
fieldsobject[]yesField definitions — see field types.
titleFieldstringnoField that names documents in lists.

Reference fields need to: [typeName], and those types must already exist — create the target type first. A list is one field of type array whose of is the single definition every item shares — an object, not a list: {"name": "authors", "title": "Authors", "type": "array", "of": {"type": "reference", "to": ["author"]}}.

update_content_type​

ParameterTypeRequiredDescription
contentTypeIdstringyesType id, or its machine name (post).
projectIdstringnoWhere a machine name is looked up. Defaults to the default project.
baseVersionnumberyesMust equal the type's latestVersion.
title · description · fields · titleFieldnoOmitted keys keep their stored value.

Changing fields mints a new schema version and goes live immediately. A stale baseVersion errors — refetch with get_content_type and retry.

delete_content_type​

ParameterTypeRequiredDescription
contentTypeIdstringyesType id, or its machine name (post).
projectIdstringnoWhere a machine name is looked up. Defaults to the default project.
forcebooleannoAlso delete every document of this type.

Refused outright while another type embeds this one — fix those fields first. Refused while documents use it unless force, which deletes those documents too, along with every schema version. Not reversible.

list_content_documents​

ParameterTypeRequiredDescription
projectIdstringnoDefaults to the default project.
typestringnoContent type machine name.
status"draft" | "published" | "changed"noFilter by state.
qstringnoSearches the type's title field (every locale of a localized one). Requires type.
missingLocalestringnoOnly documents with an empty localized field in this locale. Requires type.
limit · offsetnumbernoPaging.

Rows of a type with localized fields carry locales: per-locale coverage (complete, partial, outdated or empty, with the field paths behind it).

get_content_document​

Returns the draft and published payloads, the schema, and valid/issues — whether the draft would publish. The CRDT binaries the browser editor uses are stripped; an agent has no local document to apply them to.

create_content_document​

ParameterTypeRequiredDescription
projectIdstringnoDefaults to the default project.
typestringyesContent type machine name.
draftobjectnoField values keyed by field name.

Media is { "storageId": "…" }, references are { "documentId": "…" }. The document starts unpublished.

update_content_document​

ParameterTypeRequiredDescription
documentIdstringyesDocument id.
draftobjectyesThe fields you are changing. Others keep their values.

A partial draft is merged over the document's current draft, so changing one field leaves the rest alone. Pass a field as null to clear it. The same rule holds one level down for a localized field: { "title": { "de": "…" } } sets the German and keeps the other locales.

No version to pass: the draft merges with anyone editing live. The response carries valid and issues; a draft saves even when invalid, and publish is what refuses.

publish_content_document​

ParameterTypeRequiredDescription
documentIdstringyesDocument id.
baseVersionnumberyesThe draftVersion from get_content_document.

A mismatched baseVersion means someone edited the document since you looked. Publishing an invalid draft is refused.

unpublish_content_document​

ParameterTypeRequiredDescription
documentIdstringyesDocument id.
forcebooleannoUnpublish despite inbound published references.

Removes the live copy; the draft and the revision history stay. Anything reading published content stops seeing it immediately.

delete_content_document​

ParameterTypeRequiredDescription
documentIdstringyesDocument id.
forcebooleannoDelete despite inbound published references.

Deletes the draft, the published copy and the whole revision history. Not reversible — prefer unpublish_content_document when the intent is only to take something off a live site. Forcing past the reference guard leaves those references dangling.

list_content_revisions​

Draft saves plus publish, unpublish and restore steps, newest first.

get_content_settings​

ParameterTypeRequiredDescription
projectIdstringnoDefaults to the default project.

The project's locales (code, title, fallback, required), its default locale, and the translation settings (tone-of-voice brief, auto-translate on publish). Read it before adding a locale or translating.

update_content_settings​

ParameterTypeRequiredDescription
projectIdstringnoDefaults to the default project.
defaultLocalestringnoMust be one of locales.
localesobject[]noReplaces the whole list — read it first and resend it with the change.
translationobjectno{ brief?, autoTranslateOnPublish? }; omitted keys keep their value.

Refused when a code repeats, the default is not in the list, or a fallback points at itself, at an unknown locale, or around a loop.

translate_content_document​

ParameterTypeRequiredDescription
documentIdstringyesDocument id.
localesstring[]noTarget codes. Default: every locale but the source.
fieldsstring[]noLocalized field paths. Default: all of them.
sourceLocalestringnoDefault: the project's default locale.
mode"missing" | "outdated" | "all"noDefault outdated.

Machine-translates into the draft — nothing is published. missing fills empty values; outdated also redoes machine translations whose source text changed; all rewrites hand-written translations too, so confirm with the user first. Media fields are skipped. Costs credits, billed to the project owner; the response reports translated and skipped paths per locale and the credits spent. See Locales and translation.

create_content_delivery_token​

ParameterTypeRequiredDescription
projectIdstringnoDefaults to the default project.
namestringyesWhat the token is for.
mode"published" | "preview"noDefault published.

Requires project ownership.

:::caution The secret is returned once create_content_delivery_token returns a secret that cannot be recovered. Write it straight into the application's configuration — a CORTE_CONTENT_TOKEN environment variable, say — before doing anything else, and don't echo it back into the conversation. Use published mode for anything a browser can read; preview reads unpublished drafts and belongs on a server. :::


Brand​

Every project carries one brand: a structured identity (overview, colour tokens, typography, layout, photography direction, voice, logo rules) that renders as a DESIGN.md. Two ways to fill it: import one from a source the person already has, reviewed field by field before anything lands, or create one from a brief as three candidate directions to pick from. Everything here is per project — pass the same projectId on the start tool and its polling tool.

get_brand​

The project's brand as a structured doc plus designMd.

ParameterTypeRequiredDescription
projectIdstringnoDefaults to the account's default project.

Returns { "brand": Brand, "designMd": string, "models": [...] }. Read it before generating on-brand pictures, copy or UI — brand.doc.photography.promptSuffix is the generation-ready tail for image prompts.

get_brand_design_md​

The brand rendered as DESIGN.md text — drop it into a repo or a system prompt as-is.

ParameterTypeRequiredDescription
projectIdstringnoDefaults to the account's default project.

Returns the Markdown document as text.

update_brand​

Replace the brand doc in full.

ParameterTypeRequiredDescription
docobjectyesThe whole BrandDoc — read get_brand first and resend it with your changes.
projectIdstringnoDefaults to the account's default project.

Returns the same shape as get_brand. Unknown keys are dropped; values are validated.

import_brand​

Extract a brand from an existing source and stage it for review, waiting for the extraction when asked.

ParameterTypeRequiredDescription
kindstringnourl (default), figma, html, screenshot or designmd.
urlstringfor url / figmaThe site's public https address, or the Figma file URL or key.
figmaTokenstringnoA Figma personal access token, used for this import only and never stored. Leave it out when the project has a Figma connection.
htmlstringfor htmlThe page markup.
markdownstringfor designmdThe DESIGN.md text.
storageIdsstring[]for screenshotUp to 6 committed image storageIds of the design.
filenamestringnoA label for an html or designmd import.
schemestringnoauto (default), light or dark — which colour scheme the source shows.
projectIdstringnoThe project whose brand this is for — the owner pays.
waitSecondsnumbernoDefault 60, max 300. 0 returns immediately.

Returns { "import": BrandImport, "diff": BrandFieldChange[] }. import.status is running, ready, applied or failed; import.progress names the stage under way. Once ready, diff lists every field the source would change against the project's current brand — { path, page, label, kind, before, after } — which is what to show the person. Nothing is written to the brand until apply_brand_import.

Cost: a website or HTML import needs at least 2 credits available, a Figma or screenshot import 5; a DESIGN.md import is parsed on the spot and costs nothing. Plain-http addresses, private hosts and localhost are refused before anything is billed.

If the import is still running when the wait budget expires, it is returned in its current state — poll with get_brand_import.

Example call
{
"url": "https://acme.com/",
"projectId": "5b1e…",
"waitSeconds": 120
}
Returns
{
"import": {
"id": "d2c9…",
"projectId": "5b1e…",
"kind": "url",
"label": "https://acme.com/",
"status": "ready",
"progress": { "stage": null, "done": [{ "stage": "fetch", "detail": "3 stylesheets" }, { "stage": "model" }, { "stage": "diff", "detail": "14 changes" }] },
"extracted": { "overview": { "name": "Acme" }, "colors": { "tokens": [{ "name": "primary", "value": "#ff6600", "role": "primary" }] }, "…": "…" },
"appliedPaths": null,
"credits": 3,
"error": null,
"createdAt": "2026-09-26T14:11:55.902Z",
"appliedAt": null
},
"diff": [
{ "path": "overview.name", "page": "overview", "label": "Name", "kind": "added", "before": "", "after": "Acme" },
{ "path": "colors.tokens[primary].value", "page": "colors", "label": "primary", "kind": "added", "before": null, "after": "#ff6600" }
]
}

get_brand_import​

Fetch a brand import by id — the polling tool.

ParameterTypeRequiredDescription
importIdstringyesFrom import_brand or list_brand_imports.
projectIdstringnoThe project the import was started in.

Returns the same shape as import_brand. diff is recomputed against the brand as it is now, so a field applied earlier drops out of it.

list_brand_imports​

The project's imports, newest first.

ParameterTypeRequiredDescription
projectIdstringnoDefaults to the account's default project.
limitnumbernoPage size, default 50, max 100.
offsetnumbernoDefault 0.

Returns { "imports": BrandImport[], "total": number }.

apply_brand_import​

Make a ready import part of the brand — all of it, or the fields the person chose.

ParameterTypeRequiredDescription
importIdstringyesA ready import.
pathsstring[]nodiff[].path values to accept; omit to apply every change. Paths not in the diff are ignored.
projectIdstringnoThe project the import was started in.

Returns the same shape as get_brand: the saved brand and its DESIGN.md. Refused while the import is running or when it failed, and answers a conflict when the brand changed since the diff was computed — fetch the import again and retry. The import is marked applied with the accepted paths; what was left out stays on offer.

create_brand_directions​

Create a brand from a brief: three candidate directions, each a complete identity with a hero picture, an ad picture and first copy.

ParameterTypeRequiredDescription
brief.whatstringyesWhat the company makes or does, in a sentence or two.
brief.feelstring[]yesOne to five words people should feel.
brief.audiencestringnoWho it is for.
brief.namestringnoWorking name; leave empty to have one proposed.
brief.tasteUrlstringnoAn https site whose taste to lean toward.
brief.avoidUrlstringnoAn https site to be unlike.
projectIdstringnoThe project the brand is for — the owner pays.
waitSecondsnumbernoDefault 60, max 300. 0 returns immediately.

Returns { "session": BrandCreateSession }. session.status is running, ready, failed or accepted; while running, session.stage (brief → worlds → heroes) and session.detail say what is happening. Once ready, session.worlds[] holds the three directions: id, title, rationale, a full doc, tagline, adHeadline, adCta, postCaption, and hero / ad — each { downloadUrl, libraryUrl } for the generated pictures, the links to show the person. Nothing is saved to the project until accept_brand_direction.

Cost: needs at least 90 credits available; the model call and six pictures are billed to the project owner as they land.

Example call
{
"brief": { "what": "Hand tools for woodworkers", "audience": "Small workshops", "feel": ["precise", "warm"] },
"projectId": "5b1e…",
"waitSeconds": 0
}
Returns
{
"session": {
"id": "7a40…",
"projectId": "5b1e…",
"status": "running",
"stage": "brief",
"worlds": [],
"credits": 0,
"error": null,
"acceptedWorldId": null,
"createdAt": "2026-09-26T14:12:01.118Z"
}
}

get_brand_directions​

Fetch a brand-creation session by id — the polling tool.

ParameterTypeRequiredDescription
sessionIdstringyesFrom create_brand_directions.
projectIdstringnoThe project the session was started in.

Returns the same shape as create_brand_directions.

accept_brand_direction​

Make one direction the project's brand.

ParameterTypeRequiredDescription
sessionIdstringyesA ready session.
directionIdstringyessession.worlds[].id of the chosen direction.
projectIdstringnoThe project the session was started in.

Returns the same shape as get_brand. Only for a project whose brand is still empty — a project that already has one is changed through import_brand, which reviews every field. The direction's pictures join the brand's photography references and examples, and the session is marked accepted.


Workflows​

list_workflow_node_types​

The vocabulary for composing a workflow: every node type with what it is for, its ports, the data keys it reads, which run_workflow input key overrides it at run time, and which list_models kind its modelId accepts. Read it before create_workflow. Free, and the same for every account.

No parameters.

Returns

{
"nodeTypes": [
{
"type": "generateImage",
"label": "Generate Image",
"purpose": "Makes one image from the connected prompt, optionally guided by reference images.",
"inputs": [
{ "id": "prompt", "type": "text", "label": "Prompt", "optional": false, "multiple": false },
{ "id": "images", "type": "image", "label": "References", "optional": true, "multiple": true }
],
"outputs": [{ "id": "out", "type": "image", "label": "out" }],
"data": [
{ "key": "modelId", "type": "string", "description": "A model id from list_models (kind \"image\"). Omit for the default." },
{ "key": "aspectRatio", "type": "string", "description": "One of the model's aspectRatios." },
{ "key": "resolution", "type": "string", "description": "One of the model's resolutions." }
],
"modelKind": "image"
},
{ "type": "prompt", "label": "Prompt", "runInput": "text", "…": "…" }
],
"total": 17,
"guide": "A graph is nodes plus edges. Every node has a unique id you choose …"
}

runInput is present on the node types run_workflow can override ("text" for prompt and generateText, "storageId" for the media inputs). data[].values lists the accepted values where there is a fixed set (frame grab, social format).

create_workflow​

Save a new workflow from a graph you compose. The graph is checked before it is stored — unknown node types, edges to nodes that do not exist, handles that are not ports, a text port wired to an image port — and every problem comes back in one refusal. Saving is free; running spends credits.

ParameterTypeRequiredDescription
namestringyesWhat the pipeline does, in the person's words.
graphobjectyes{ nodes, edges } — see below.
projectIdstringnoA shared project to save it in; omit for the default project.

graph.nodes[]: { id, type, data?, x?, y? } — id is yours to choose and is what run_workflow inputs are keyed by; type and data come from list_workflow_node_types; omit x/y to have the nodes laid out left-to-right by dependency. graph.edges[]: { source, sourceHandle?, target, targetHandle?, id? } — handles are port ids and default to the first port on each side; id is generated when omitted.

Returns the saved workflow and the graph as stored.

Example call
{
"name": "Fox hero",
"graph": {
"nodes": [
{ "id": "concept", "type": "prompt", "data": { "text": "a fox in tall grass, dawn" } },
{ "id": "hero", "type": "generateImage", "data": {} },
{ "id": "final", "type": "output", "data": {} }
],
"edges": [
{ "source": "concept", "sourceHandle": "out", "target": "hero", "targetHandle": "prompt" },
{ "source": "hero", "sourceHandle": "out", "target": "final", "targetHandle": "image" }
]
}
}
{
"workflow": { "id": "7c93b1de-…", "name": "Fox hero", "nodeCount": 3, "projectId": "1b6d4f90-…", "createdAt": "…", "updatedAt": "…", "workflowUrl": "https://app.corte.so/workflows/7c93b1de-…" },
"graph": {
"nodes": [
{ "id": "concept", "type": "prompt", "label": "Prompt", "x": 0, "y": 0, "data": { "text": "a fox in tall grass, dawn" } },
{ "id": "hero", "type": "generateImage", "label": "Generate Image", "x": 320, "y": 0, "data": {} },
{ "id": "final", "type": "output", "label": "Output", "x": 640, "y": 0, "data": {} }
],
"edges": [
{ "id": "e1-concept-hero", "source": "concept", "sourceHandle": "out", "target": "hero", "targetHandle": "prompt" },
{ "id": "e2-hero-final", "source": "hero", "sourceHandle": "out", "target": "final", "targetHandle": "image" }
]
}
}

Validation matches POST /v1/workflows, plus the port check above.

list_workflows​

List workflows in a project, most recently updated first — the account's default project unless projectId is given. Scoping matches GET /v1/workflows: every member's workflows in a shared project, not just the agent's.

ParameterTypeRequiredDescription
projectIdstringnoA shared project to list; omit for the default project.

Returns

{
"workflows": [
{ "id": "7c93b1de-…", "name": "Product Reshoot Pipeline", "nodeCount": 6, "updatedAt": "2026-07-27T09:41:22.006Z", "projectId": "1b6d4f90-…" }
],
"total": 12
}

get_workflow​

Inspect a workflow and discover which node ids accept run inputs. This is the tool to call before run_workflow.

ParameterTypeRequiredDescription
workflowIdstringyesWorkflow id. Same access rule as REST — project members included.

Returns

{
"workflow": { "id": "7c93b1de-…", "name": "Product Reshoot Pipeline", "nodeCount": 6 },
"nodes": [
{ "id": "productPhoto", "type": "imageInput" },
{ "id": "concept", "type": "prompt" },
{ "id": "reshoot", "type": "generateImage" },
{ "id": "final", "type": "output" }
],
"inputNodes": [
{ "nodeId": "productPhoto", "type": "imageInput", "takes": "storageId", "current": "b0e7c9a1-…" },
{ "nodeId": "concept", "type": "prompt", "takes": "text", "current": "Alpine trail-running capsule, cold dawn light" }
]
}

inputNodes is the useful part: takes says which key to send ("text" or "storageId"), and current shows the saved value that will be used if you omit the node.

run_workflow​

Execute a workflow server-side with optional per-node overrides, and wait.

ParameterTypeRequiredDescription
workflowIdstringyesWorkflow id.
inputsobjectnoMap of node id → { text } or { storageId }. Omitted nodes keep saved values.
projectIdstringnoBilling context for a personal workflow run inside a shared project.
waitSecondsnumbernoDefault 120, max 600. 0 returns immediately.

Returns { "run": Run } with nodeResults populated as far as the run got.

Example call
{
"workflowId": "7c93b1de-2a48-4f60-91b7-5d0e6a2c8f13",
"inputs": {
"concept": { "text": "Alpine trail-running capsule, cold dawn light" },
"productPhoto": { "storageId": "b0e7c9a1-4d52-4f0e-8c31-77a2e5d9f6b8" }
},
"waitSeconds": 300
}

Input validation matches the REST endpoint exactly — see POST /v1/workflows/:id/execute.

get_workflow_run​

Poll a run for status and per-node results.

ParameterTypeRequiredDescription
workflowIdstringyesThe workflow the run belongs to.
runIdstringyesRun id. Visibility matches REST: all members see a project workflow's runs.

Returns { "run": Run } — status plus node results with freshly signed media URLs. nodeResults grows between polls while the run is running.

update_workflow​

Rename a workflow, replace its graph, or both. Omitted fields are unchanged. A graph you send replaces the whole stored graph, so read it first with get_workflow and send the complete new one, keeping the ids of nodes people already refer to in run inputs.

ParameterTypeRequiredDescription
workflowIdstringyesWorkflow id. Needs write access: the creator, or any member of its shared project.
namestringnoNew name. Cannot be empty.
graphobjectnoComplete replacement graph, same shape and checks as create_workflow.

Returns the same shape as create_workflow.

delete_workflow​

Delete a workflow and its run history. Cannot be undone; media its runs produced stays in the library. Only the creator or the project owner may delete one.

ParameterTypeRequiredDescription
workflowIdstringyesWorkflow id.

Returns { "ok": true, "deleted": { "id": "…", "name": "…" } }.

list_workflow_runs​

A workflow's run history, newest first. Visibility matches the app: every member sees every run of a shared-project workflow; a personal workflow shows only the caller's runs.

ParameterTypeRequiredDescription
workflowIdstringyesWorkflow id.
limitnumbernoDefault 20, max 100.
offsetnumbernoDefault 0.

Returns

{
"workflow": { "id": "7c93b1de-…", "name": "Fox hero", "workflowUrl": "https://app.corte.so/workflows/7c93b1de-…" },
"runs": [
{ "id": "9f1e…", "status": "complete", "createdAt": "…", "completedAt": "…", "userId": "…", "nodesFinished": 3 }
],
"total": 1,
"limit": 20,
"offset": 0
}

status is one of running, complete, error, stopped. Fetch a run's per-node results with get_workflow_run.

cancel_workflow_run​

Stop a run that is still running. Finished steps keep their results and stop being billed further. The run settles as stopped within a few seconds, so the returned row may still read running; get_workflow_run shows it settle. Stopping a settled run changes nothing. Only the person who started a run can stop it.

ParameterTypeRequiredDescription
workflowIdstringyesThe workflow the run belongs to.
runIdstringyesRun id.

Returns { "run": { "id", "status", "createdAt", "completedAt", "userId", "nodesFinished" } }.


Build​

Apps Corte builds, hosts and deploys. Reading them needs build:read; asking an app's coding agent for a change, or deploying one, needs build:rw.

These are the tools that let an agent maintain a website on its own — read what the site is doing, ask for a change, publish it.

list_build_apps​

The Build apps the account can work in — its own, and those in projects it has been invited to.

ParameterTypeRequiredDescription
limit · offsetnumbernoPaging. Default limit 25.

Returns { items, total, limit, offset }. Each item carries the id the other Build tools take, plus name, githubRepo, defaultBranch, deployTarget and access — manage for apps the account owns or created (it may also change the address, secrets and database, or delete the app) and edit for apps shared with it as a contributor (chat, build and deploy).

get_build_app​

One app's current state, in a single call.

ParameterTypeRequiredDescription
appIdstringyesFrom list_build_apps.

Returns { app, session, recentCheckpoints, checkpointTotal, recentDeploys, deployTotal }.

  • session is the app's live or parked builder thread, or null.
  • A checkpoint is a commit the app's coding agent made. Newest first; its sha is what trigger_deploy takes.
  • A deploy carries status and, once live, the public url.

The two lists are capped at 10 — a summary to act on, not a paging surface.

create_build_app​

Create an app and hand its builder the brief — the "New app" step in one call. The brief is the first message on the app's thread, exactly as if the owner had typed it.

ParameterTypeRequiredDescription
namestringyesThe app's name — becomes its public subdomain (<slug>.corte.sh), so name the product, not the request.
promptstringyesThe brief the builder starts from. Answer a skill's interview inside it (see get_skill) and the builder starts without asking.
projectIdstringnoOwn the app in a shared project; defaults to the account's default project.

Returns { app, sessionId, messageId, publicUrl }. Starts a sandbox and spends the owner's credits. Returns as soon as the brief is queued — follow with get_build_thread.

update_build_app​

Rename an app and/or move its public address.

ParameterTypeRequiredDescription
appIdstringyesThe app.
namestringnoNew display name.
slugstringnoNew public subdomain (<slug>.corte.sh). Lowercase letters, digits, hyphens.

Returns { app, publicUrl }. The old address stops serving once the new one is live.

get_build_thread​

What the builder has said and done — the reply side of send_build_message and create_build_app.

ParameterTypeRequiredDescription
appIdstringyesThe app.
sessionIdstringnoA specific thread; defaults to the newest.
sincenumbernoOnly events after this id — pass back the previous call's lastEventId.
waitSecondsnumbernoBlock up to this long (max 300) for a reply, a question, a checkpoint or the thread ending. Default 0.

Returns { session, lastEventId, replies, pendingQuestion, checkpoints, errors, settled }. session.status is queued | provisioning | running | awaiting_input | completed | failed | cancelled. A pendingQuestion is answered with send_build_message.

list_skills · get_skill​

The platform skills a builder can run — packaged procedures for a kind of site or section (scroll-craft: a premium scroll-driven landing page where every act behaves differently, with one engineered peak; scroll-world: one continuous camera flight through generated scenes). list_skills takes no parameters and returns { skills: [{ name, description }] }; get_skill takes name and returns { name, description, interview, body }, where interview is the list of questions a brief should answer up front. Mention the skill by name in the brief.

send_build_message​

Ask an app's coding agent to change the app, in plain language, exactly as its owner would.

ParameterTypeRequiredDescription
appIdstringyesThe app to change.
textstringyesWhat you want changed.

Returns { appId, sessionId, messageId, startedSession }.

:::caution Accepted is not done This returns as soon as the request is queued, not when the change exists. The builder works asynchronously in its own sandbox. Poll get_build_app — a new checkpoint means it committed something — and remember that a committed change is not live until it is deployed. :::

If the app has no live thread, one is opened, which starts a sandbox and spends the owner's credits. startedSession tells you which happened.

trigger_deploy​

Publish an app to its live URL.

ParameterTypeRequiredDescription
appIdstringyesThe app to deploy.
checkpointShastringnoDefaults to the newest checkpoint.

Returns the deploy, which starts queued. Poll get_build_app for how it ended and the URL it went to.

This makes changes live for real visitors.

:::note Unknown, not forbidden An app the account does not own reads as unknown, for the same reason assets do: a "forbidden" answer would confirm the id exists. :::


Agents​

Agents are standing automations: a prompt that runs on a schedule, on a content event, or on demand, with its own scopes. Reading them needs agents:read; creating, editing or running one needs agents:rw.

These tools exist so an assistant can author an automation for someone, not just perform a task. Agents are owner-only — only a project's owner sees or manages its agents — and a project holds at most 20.

:::warning An agent is a credential Creating one hands out standing, unattended access to the account, exercised on a schedule until the agent is disabled or deleted. Two rules bound it, both enforced on this path exactly as on REST:

  • An agent can never exceed its creator. A key scoped to content:rw cannot create an agent holding build:rw. The request is refused, not trimmed.
  • The account area is refused outright, at any level and to anyone. It reaches API key creation, so a run could mint itself a credential that outlives the agent — and revocability is the whole point of a per-run credential. :::

list_agents​

The agents this account has created.

ParameterTypeRequiredDescription
projectIdstringnoOne project's agents. Omit for every agent you created.
limit · offsetnumbernoPaging. Default limit 50.

Returns { agents, total, limit, offset }.

get_agent​

One agent in full — its prompt, scopes, schedule, events, and lastRunAt / nextRunAt.

ParameterTypeRequiredDescription
agentIdstringyesFrom list_agents.

Returns { agent }. Read this before update_agent: an update replaces the fields it is given, so you need to know what is already there.

list_agent_recipes​

The curated agent recipes — ready-made automations, each with a prompt and suggested scopes, schedule, time zone and model. Takes no parameters.

Returns { recipes }. Read it before create_agent: when one matches what the user described, create the agent with that recipe's id as recipeId so it tracks improvements to the recipe automatically.

A suggestion is not a grant — suggestedScopes are still bounded by what the caller holds, so a recipe can never widen what you can hand out.

create_agent​

Create an agent.

ParameterTypeRequiredDescription
namestringyesShort label the owner will recognise.
promptstringyesThe standing instruction, run unattended. Max 20,000 characters.
scopesstring[]yesWhat it may do — "area:read" or "area:rw". See below.
projectIdstringnoDefaults to the account's default project.
descriptionstringnoOne line on what it's for.
scheduleCronstringno5-field cron. Omit for a manual-only agent.
scheduleTzstringnoIANA zone name for the cron. Defaults to UTC.
eventsstring[]noContent events that trigger a run.
modelIdstringnoDefaults to Auto.
recipeIdstringnoTrack a recipe from list_agent_recipes.
enabledbooleannoDefault true.

Returns { agent }, including the id the other agent tools take.

Scopes are the same vocabulary as API keys: content, media, integrations, generation, workflows, cuts, projects, plugins, community, build and agents, each :read or :rw. rw covers reading too. account is not grantable to an agent.

scheduleCron is standard 5-field cron — minute, hour, day-of-month, month, day-of-week. 0 9 * * 1-5 is 09:00 every weekday. scheduleTz must be an IANA zone name (America/New_York, Europe/Lisbon) — an abbreviation like EST or an offset is refused, because only a zone name follows daylight saving. Omit both and the agent is manual-only: it fires only via run_agent or an event.

events accepts exactly document.publish, document.unpublish and document.delete — a run per matching change in the agent's project. Draft saves never trigger anything.

update_agent​

Edit an agent. Only the fields you send change.

ParameterTypeRequiredDescription
agentIdstringyesThe agent to edit.
everything from create_agentnoExcept projectId, which cannot change.

Returns { agent }.

Two fields replace rather than merge: scopes and events. Send the full list you want the agent to end up with, or it loses what you left out — and the replacement is bounded by what you hold, exactly as at creation.

  • scheduleCron: null removes the schedule and leaves the agent manual-only.
  • enabled: false pauses it without deleting it. Re-enabling resumes from the next slot; missed slots are not made up.

run_agent​

Start a run now, without waiting for the schedule.

ParameterTypeRequiredDescription
agentIdstringyesThe agent to run.
idempotencyKeystringnoRepeat the same key to retry safely.

Returns { run } with a status of queued, launching or running.

:::caution Queued is not finished This returns as soon as the run is queued. The agent works asynchronously and spends the owner's credits, so a successful call is not the work being done. get_agent shows lastRunAt once a run lands; list_agent_runs is the run feed, and each run carries a link into the app. :::

A run is refused with a reason when the account is already at its concurrent-run limit. A disabled agent still runs when asked directly.

list_agent_runs​

The run history of one agent — what each run was about, its status, what it made, and what it cost.

ParameterTypeRequiredDescription
agentIdstringyesFrom list_agents.
threadRunIdstringnoA top-level run id: return its replies, oldest first, instead of the channel.
allbooleannoEvery run including replies, newest first.
limit · offsetnumbernoPaging. Default limit 50.

Returns { runs, total, limit, offset }. By default the channel view: top-level runs newest first, each with status, statusDetail, subject, report, outputs, creditsSpent (everything the run cost, the media it generated or transcribed included), mediaCreditsSpent (that media share), replyCount, needsInput and a url that opens the run in the app. Any member of the agent's project can read it.

status is queued, launching or running while in flight, then one of succeeded, failed, canceled, timed_out or skipped. A run with needsInput: true ended on a question — answer it with reply_to_agent_run.

cancel_agent_run​

Stop a run that is still in flight.

ParameterTypeRequiredDescription
agentIdstringyesThe agent the run belongs to.
runIdstringyesFrom list_agent_runs or run_agent.

Returns { run } with status: "canceled". What the run already did stays done. Only a queued, launching or running run can be stopped; asking to cancel a finished one is refused with a conflict naming the state it is already in. Owner-only.

list_agent_outputs​

Everything an agent has made across all its runs, newest first.

ParameterTypeRequiredDescription
agentIdstringyesFrom list_agents.
limit · offsetnumbernoPaging. Default limit 50.

Returns { outputs, total, limit, offset }. Each output has a kind (content_document or social_post), the refId of that thing, its title, the action the run took (created, updated or published), the runId and when. Open a document with get_content_document; a post with list_growth_posts. Any member of the agent's project can read it.

{
"outputs": [
{
"id": "aart_…",
"runId": "arun_…",
"kind": "content_document",
"refId": "doc_…",
"contentType": "post",
"title": "What shipped this week",
"action": "published",
"at": "2026-09-26T09:02:11.000Z"
}
],
"total": 1,
"limit": 50,
"offset": 0
}

delete_agent​

Delete an agent for good.

ParameterTypeRequiredDescription
agentIdstringyesThe agent to delete.

Returns { ok: true }. Removes the definition, schedule, event subscriptions, standing credential and the whole run history including replies and transcripts; a run still in flight is stopped first. What the agent made is not touched. There is no undo and no archive — to stop an agent without losing it, update_agent with enabled: false. Owner-only.

list_agent_memories​

The notes an agent is briefed with before every run, strongest first.

ParameterTypeRequiredDescription
agentIdstringyesFrom list_agents.
limit · offsetnumbernoPaging. Default limit 50.

Returns { memories, total, limit, offset }. Each note has a source — run for something the agent learned from its own earlier runs, which fades with age, or user for an instruction its owner wrote, held at full strength until deleted — plus content, effectiveConfidence and injected (whether a run would still be shown it). Faded notes are still listed so they can be found and deleted. Owner-only.

add_agent_memory​

Give an agent a standing instruction — "from now on, …".

ParameterTypeRequiredDescription
agentIdstringyesThe agent to instruct.
contentstringyesOne short statement. Max 400 characters.

Returns { memory }, including the id that delete_agent_memory takes. The note never fades. An agent holds at most 50 notes; past that the weakest learned notes make room first, and what a person wrote always outranks what the agent learned by itself. Owner-only.

delete_agent_memory​

Remove one note from an agent.

ParameterTypeRequiredDescription
agentIdstringyesThe agent the note belongs to.
memoryIdstringyesFrom list_agent_memories.

Returns { ok: true }. This is how a person corrects an agent that learned something wrong. The agent may learn the same thing again from a later run if the reason it learned it is still there. Owner-only.


Integrations​

The outside services a project can act through — a Shopify store, an ad account, a Notion workspace. A connection belongs to a project: every member can use it, and it never follows the person who connected it into their other projects. Reading needs integrations:read.

list_integrations​

What this project is connected to, and what it could still connect.

ParameterTypeRequiredDescription
projectIdstringnoDefaults to the project this credential acts in, else the default project.

Returns { connected, available }. connected lists live connections — provider id, label, and a status of connected, expiring or expired (never credentials); the provider-specific tools like shopify_list_assets only work for these, and an expiring one needs reconnecting before its tools are reliable. available lists the providers the platform supports that the project has not connected, with connectable saying whether this server can connect them. When a request needs data no connected service provides, look here before calling it impossible — the right answer is usually to suggest connecting the provider that covers it.

Connections are made in the app, in Settings → Integrations, and only by a person — there is no tool that connects a service.


Growth​

Growth is the social publish queue: a channel is a connected social account in a project, a post is a row that moves draft → scheduled → published (or failed), and a campaign groups one story's posts. A post a tool drafts lands in pending_review, and only a person — in Growth's review lane — moves it to scheduled. There is deliberately no approve or schedule tool: an agent can draft, fix, decline, pull back and retry, and never put words in public that nobody read. Every post carries a postUrl that opens it in the app.

list_growth_channels​

The social accounts a project can publish to — one channel per connected platform.

ParameterTypeRequiredDescription
projectIdstringnoDefaults to your default project.
limitnumbernoDefault 50, max 100.
offsetnumbernoDefault 0.

Returns { "channels": GrowthChannel[], "total", "limit", "offset" }. Each channel has the id a post needs, its provider, handle, displayName and status (active, or disconnected when the account link is gone). A social account connected in Settings appears here on the next call.

create_growth_post​

Draft a social post into the queue. By default it lands in pending_review, where the person approves, edits or declines it.

ParameterTypeRequiredDescription
channelIdstringyesFrom list_growth_channels.
textstringyesThe post, written for that platform. Checked against its limits and refused with the reason when it would not post. Links become tracked links.
mediaarraynoUp to 4 { storageId, alt? } — library asset ids; the platform gets a published copy.
publishAtstringnoISO 8601 suggestion; the reviewer confirms or changes it.
campaignIdstringnoA campaign in the same project.
statestringnopending_review (default) or draft — parked, not up for review. Nothing a tool creates is scheduled.
settingsobjectnoPlatform-specific options the channel's platform understands.
projectIdstringnoThe channel's project; defaults to your default project.

Returns { "post": GrowthPost & { postUrl } }.

Example call
{
"channelId": "6f1e9c4a-2b7d-4e58-9a13-c0d4e8b2f715",
"text": "Spring lookbook is live. Every look, one link: https://corte.so/lookbook",
"media": [{ "storageId": "c74a1e08-9b6f-4d13-a205-e8f30b7c9d61", "alt": "Model in the linen set on a stone terrace" }],
"publishAt": "2026-10-02T09:00:00Z",
"campaignId": "b3a0d9e2-71c5-4f0b-8e2a-5d6f7c8a9b01"
}
Returns
{
"post": {
"id": "0c9d2e7f-4a1b-4c3d-9e8f-1a2b3c4d5e6f",
"projectId": "…",
"channelId": "6f1e9c4a-2b7d-4e58-9a13-c0d4e8b2f715",
"campaignId": "b3a0d9e2-71c5-4f0b-8e2a-5d6f7c8a9b01",
"state": "pending_review",
"publishAt": "2026-10-02T09:00:00.000Z",
"body": { "text": "Spring lookbook is live. Every look, one link: https://go.corte.so/x7Kq2", "media": [ { "storageId": "…", "url": "https://media.corte.so/pub/…", "alt": "…" } ] },
"source": { "actor": "human" },
"platformRef": null,
"platformUrl": null,
"attempts": 0,
"lastError": null,
"createdAt": "2026-09-26T14:10:02.118Z",
"publishedAt": null,
"postUrl": "https://app.corte.so/growth/calendar/0c9d2e7f-4a1b-4c3d-9e8f-1a2b3c4d5e6f"
}
}

list_growth_posts​

The queue across every state, with per-post click counts where links are tracked.

ParameterTypeRequiredDescription
projectIdstringnoDefaults to your default project.
statestringnoOne of draft, pending_review, scheduled, publishing, pending_platform, published, failed, declined.
campaignIdstringnoOnly that campaign's posts.
limitnumbernoDefault 50, max 100.
offsetnumbernoDefault 0.

Returns { "posts": (GrowthPost & { clicks? })[], "total", "limit", "offset" }. clicks absent means untracked, not zero.

get_growth_post​

One post: text, media, channel, campaign, state, publish time, the live URL once it is out, and lastError when a send failed.

ParameterTypeRequiredDescription
postIdstringyesPost id. Visible to the project owner.

Returns { "post": GrowthPost & { postUrl } }.

update_growth_post​

Change a post that is still a draft or awaiting review. Omitted fields stay as they are; text and media each replace the whole value.

ParameterTypeRequiredDescription
postIdstringyes
textstringnoNew text, checked against the channel's platform.
mediaarraynoNew media set, { storageId, alt? }[].
settingsobjectnoPlatform-specific options.
publishAtstring · nullnoNew suggested time; null clears it.
channelIdstringnoMove it to another channel in the same project.
campaignIdstring · nullnonull detaches it from its campaign.

Returns { "post": GrowthPost & { postUrl } }. A scheduled post answers conflict — a person approved those words. Pull it back with unschedule_growth_post, edit, and the person schedules it again in Growth.

decline_growth_post​

Decline a post awaiting review. It leaves the lane and stays as history in declined, never published.

ParameterTypeRequiredDescription
postIdstringyesMust be pending_review.

Returns { "post": GrowthPost & { postUrl } }.

unschedule_growth_post​

Pull a scheduled post back to draft before it goes out. It keeps its text, media and time and will not publish until a person schedules it again.

ParameterTypeRequiredDescription
postIdstringyesMust be scheduled. Once the send has started the answer is conflict.

Returns { "post": GrowthPost & { postUrl } }.

retry_growth_post​

Send a failed post again — the words a person already approved. Read lastError first; a retry only helps once the cause is fixed.

ParameterTypeRequiredDescription
postIdstringyesMust be failed.
publishAtstringnoISO 8601; omit to send as soon as possible.

Returns { "post": GrowthPost & { postUrl } }.

delete_growth_post​

Remove a post from the queue for good. A scheduled post is pulled back first; one whose send is in progress cannot be deleted. Deleting a published post removes Corte's record, not the post on the platform.

ParameterTypeRequiredDescription
postIdstringyes

Returns { "ok": true, "postId" }.

get_growth_post_analytics​

How one post landed: clicks on its tracked links and, where the platform reports it, its own engagement.

ParameterTypeRequiredDescription
postIdstringyes

Returns { "enabled", "ok", "clicks", "visitors", "byDay": [{ day, clicks }], "metrics": { likes?, comments?, shares?, views? } | null, "metricsAt" }. enabled false means this deployment does not track links; ok false means the click read failed — neither is zero. metrics is null until the post is live or when the platform has no readable engagement, and a field missing inside it means that platform does not report it.

list_growth_campaigns​

The project's campaigns — each groups one story's posts.

ParameterTypeRequiredDescription
projectIdstringnoDefaults to your default project.
limitnumbernoDefault 50, max 100.
offsetnumbernoDefault 0.

Returns { "campaigns": GrowthCampaign[], "total", "limit", "offset" }.

create_growth_campaign​

Start a campaign to group a set of posts.

ParameterTypeRequiredDescription
namestringyesThe story it tells, up to 200 characters.
descriptionstringnoUp to 2,000 characters.
contentDocumentIdstringnoThe Content document it promotes, if any.
projectIdstringnoDefaults to your default project.

Returns { "campaign": GrowthCampaign } — pass its id as campaignId when drafting posts.

get_growth_campaign​

One campaign with every post in it and its combined clicks.

ParameterTypeRequiredDescription
campaignIdstringyes

Returns { "campaign", "posts": (GrowthPost & { postUrl })[], "analytics": { enabled, ok, clicks, visitors } }. The same flag rules as get_growth_post_analytics apply.

update_growth_campaign​

Rename a campaign or change its description.

ParameterTypeRequiredDescription
campaignIdstringyes
namestringno
descriptionstring · nullnonull clears it.

Returns { "campaign": GrowthCampaign }.

delete_growth_campaign​

Delete a campaign. Its posts survive, ungrouped — nothing scheduled is affected.

ParameterTypeRequiredDescription
campaignIdstringyes

Returns { "ok": true, "campaignId" }.


Ads​

Your store's products as ads: every fitting layout at every size, in one look, priced live from the catalog, approved by you and published to Meta, Google and Pinterest through the ad set's catalog feed. An agent can do everything except approve — that's update_ad, which the in-app agents are never given.

The flow: list_catalog_products (after sync_catalog if it's empty) → prepare_catalog_product for any product whose photos aren't ready → create_ad_set with styleId: "brand" and the sizes you want → generate_catalog_ads once per product. Afterwards, redesign_ad reworks one ad from an instruction, rewrite_product_copy gives a product new words, and refresh_ad_set brings the set up to date when the look, photos or offer change. Ads someone edited are left alone by the last two unless includeEdited is set.

ToolWhat it takesWhat it returns
list_catalog_productsprojectId, q?, limit?, offset?Products with prices, photos and prepared
sync_catalogprojectIdHow many products were imported and updated
prepare_catalog_productprojectId, productId, scene?The product with its prepared photos (credits per photo)
list_ad_layouts—Layouts (id, purpose, what a product needs) and house styles
create_ad_setprojectId, name, styleId, productIds, sizes?, layoutIds?, designsPerProduct?, offer?The ad set, with its feedUrl
list_ad_setsprojectIdAd sets with counts, feedUrl and refresh (what a refresh would change)
generate_catalog_adsprojectId, adSetId, productId, layouts?, designs?The new ads, skipped layouts and why, credits spent
list_adsprojectId, adSetId, productId?, status?Ads with image, layout, status, edited and checker issues
update_adprojectId, adId, status?, doc?The ad
redesign_adprojectId, adId, instruction, mode?The ad and one sentence on what changed (credits)
refresh_ad_setprojectId, adSetId, productId?, includeEdited?Rebuilt and added ads, removed and skipped counts, remaining (free)
rewrite_product_copyprojectId, adSetId, productId, includeEdited?The new copy, the ads it changed, credits spent

The canonical agent flow​

Ask an agent for "run my Product Reshoot workflow on this photo" and a well-behaved run looks like:

  1. upload_asset { url } → a committed storageId.
  2. get_workflow { workflowId } → read inputNodes to learn the node ids and what each one takes.
  3. run_workflow { workflowId, inputs: { productPhoto: { storageId }, concept: { text } } } → waits up to 120s by default.
  4. get_workflow_run if step 3 returned still-running → outputs land in run.nodeResults[*].output.value, with storageId alongside for chaining.

For one-shot generation, the flow collapses to list_models → generate, or list_applications → run_application.

Tool index​

ToolWaits?REST equivalent
get_credits—GET /v1/account
list_projects—GET /v1/projects
get_account—GET /v1/account
get_usage—GET /v1/usage
list_plans—GET /v1/plans
create_topup_checkout—POST /v1/billing/topup
create_subscription_checkout—POST /v1/billing/checkout
create_billing_portal_link—POST /v1/billing/portal
get_project—GET /v1/projects/:id
create_project—POST /v1/projects
update_project—PATCH /v1/projects/:id
invite_to_project—POST /v1/projects/:id/invites
cancel_project_invite—DELETE /v1/projects/:id/invites/:inviteId
remove_project_member—DELETE /v1/projects/:id/members/:userId
get_project_usage—GET /v1/projects/:id/usage
list_models—GET /v1/models
generate60s / max 300POST /v1/generations
get_generation—GET /v1/generations/:id
transcribe_asset120s / max 300POST /v1/transcriptions
get_transcription—GET /v1/transcriptions/:id + result
list_assets—GET /v1/projects/:id/storage
get_asset—GET /v1/storage/:storageId
upload_asset—POST /v1/uploads/import
list_applications—GET /v1/applications
run_application60s / max 300POST /v1/applications/:id/run
list_content_types—GET /v1/projects/:id/content/types
get_content_type—GET /v1/content/types/:id
create_content_type—POST /v1/projects/:id/content/types
update_content_type—PUT /v1/content/types/:id
delete_content_type—DELETE /v1/content/types/:id
list_content_documents—GET /v1/projects/:id/content/documents
get_content_document—GET /v1/content/documents/:id
create_content_document—POST /v1/projects/:id/content/documents
update_content_document—PUT /v1/content/documents/:id
publish_content_document—POST /v1/content/documents/:id/publish
unpublish_content_document—POST /v1/content/documents/:id/unpublish
delete_content_document—DELETE /v1/content/documents/:id
list_content_revisions—GET /v1/content/documents/:id/revisions
create_content_delivery_token—POST /v1/projects/:id/content/tokens
get_brand—GET /v1/projects/:id/brand
get_brand_design_md—GET /v1/projects/:id/brand/design.md
update_brand—PUT /v1/projects/:id/brand
import_brand60s / max 300POST /v1/projects/:id/brand/imports
get_brand_import—GET /v1/projects/:id/brand/imports/:importId
list_brand_imports—GET /v1/projects/:id/brand/imports
apply_brand_import—POST /v1/projects/:id/brand/imports/:importId/apply
create_brand_directions60s / max 300POST /v1/projects/:id/brand/create
get_brand_directions—GET /v1/projects/:id/brand/create/:sessionId
accept_brand_direction—POST /v1/projects/:id/brand/create/:sessionId/accept
list_workflow_node_types—none — the vocabulary the canvas is built on
create_workflow—POST /v1/workflows
list_workflows—GET /v1/workflows
get_workflow—GET /v1/workflows/:id
run_workflow120s / max 600POST /v1/workflows/:id/execute
get_workflow_run—GET /v1/workflows/:id/runs/:runId
update_workflow—PATCH /v1/workflows/:id
delete_workflow—DELETE /v1/workflows/:id
list_workflow_runs—GET /v1/workflows/:id/runs
cancel_workflow_run—POST /v1/workflows/:id/runs/:runId/stop
list_build_apps—none — Build's REST is not public API
get_build_app—none — Build's REST is not public API
send_build_message—none — Build's REST is not public API
trigger_deploy—none — Build's REST is not public API
create_build_app—none — Build's REST is not public API
get_build_thread—none — Build's REST is not public API
update_build_app—none — Build's REST is not public API
list_skills · get_skill—none — platform metadata
list_agents—GET /v1/agents
get_agent—GET /v1/agents/:id
list_agent_recipes—GET /v1/agent-recipes
create_agent—POST /v1/agents
update_agent—PATCH /v1/agents/:id
run_agent—POST /v1/agents/:id/run
list_agent_runs—GET /v1/agents/:id/runs
get_agent_run—GET /v1/agents/:id/runs/:runId
reply_to_agent_run—POST /v1/agents/:id/runs/:runId/reply
cancel_agent_run—POST /v1/agents/:id/runs/:runId/cancel
list_agent_outputs—GET /v1/agents/:id/outputs
delete_agent—DELETE /v1/agents/:id
list_agent_memories—GET /v1/agents/:id/memories
add_agent_memory—POST /v1/agents/:id/memories
delete_agent_memory—DELETE /v1/agents/:id/memories/:memoryId
list_integrations—GET /v1/integrations + /catalog
list_growth_channels—GET /v1/growth/channels
create_growth_post—POST /v1/growth/posts (state pending_review or draft only)
list_growth_posts—GET /v1/growth/posts + click counts
get_growth_post—GET /v1/growth/posts/:id
update_growth_post—PATCH /v1/growth/posts/:id (draft and pending_review only)
decline_growth_post—POST /v1/growth/posts/:id/decline
unschedule_growth_post—POST /v1/growth/posts/:id/unschedule
retry_growth_post—POST /v1/growth/posts/:id/retry
delete_growth_post—DELETE /v1/growth/posts/:id
get_growth_post_analytics—GET /v1/growth/posts/:id/analytics
list_growth_campaigns—GET /v1/growth/campaigns
create_growth_campaign—POST /v1/growth/campaigns
get_growth_campaign—GET /v1/growth/campaigns/:id
update_growth_campaign—PATCH /v1/growth/campaigns/:id
delete_growth_campaign—DELETE /v1/growth/campaigns/:id
list_catalog_products—GET /v1/projects/:projectId/catalog/products
sync_catalog—POST /v1/projects/:projectId/catalog/sync
prepare_catalog_product—POST /v1/projects/:projectId/catalog/products/:id/prepare
list_ad_layouts—GET /v1/creative/layouts
create_ad_set—POST /v1/projects/:projectId/ad-sets
list_ad_sets—GET /v1/projects/:projectId/ad-sets
generate_catalog_ads—POST /v1/projects/:projectId/ad-sets/:id/generate
list_ads—GET /v1/projects/:projectId/ad-sets/:id/ads
update_ad—PATCH /v1/projects/:projectId/ads/:adId
redesign_ad—POST /v1/projects/:projectId/ads/:adId/redesign
refresh_ad_set—POST /v1/projects/:projectId/ad-sets/:id/refresh
rewrite_product_copy—POST /v1/projects/:projectId/ad-sets/:id/products/:productId/copy

The Build tools and list_workflow_node_types are the only ones with no REST counterpart on api.corte.so: Build runs in its own cluster, and MCP is how it is reached from outside the app; the node-type vocabulary is what the canvas is built on.

Not exposed as tools​

File deletion, storage listing, API key management, and the Content delivery, webhook and GraphQL endpoints have no MCP tools — they're REST-only. An agent that needs them should call the REST API directly with the same key.

A few things stay a person's alone, in the app, and have no tool on purpose: approving or scheduling a Growth post and adding or removing a social channel; connecting an integration; deleting a project, handing it to another member or moving it between organisations.