Verbatim API and MCP server

Create a personal API key in Settings → API keys (https://verbatimcoach.com/dashboard/settings#api-keys). Send it as `Authorization: Bearer vb_...`. Keys are shown once, stored only as a SHA-256 hash, and can be revoked at any time. Create a free account to get started.

No. Each successful generate call uses one draft from the same monthly allowance as the web app: Free 3 drafts per calendar month (UTC), Pro $29/mo unlimited. Failed calls are not counted.

12 generate calls per 10 minutes per account (shared with the web app); 120 read calls per minute. Over the limit: HTTP 429 with Retry-After.

POST /api/v1/generate is synchronous and returns the finished draft, usually in 20 to 90 seconds. Use a client timeout of at least 120 seconds.

• POST /api/v1/generate: Create a draft. Body: {"input": url-or-text, "sourceType"?: "article"|"youtube"|"text", "formats"?: ["thread","carousel","newsletter","hooks"]}. Returns 201 with the result and updated usage.
• GET /api/v1/generations/{id}: Fetch a draft with all outputs.
• GET /api/v1/generations?limit=20: List recent drafts (1 to 50).
• GET /api/v1/generations/{id}/pdf: Download the LinkedIn carousel PDF.
• GET /api/v1/me: Plan, drafts used and remaining this month.
• POST /api/mcp: MCP server (Streamable HTTP, JSON-RPC). Tools: generate_content, get_generation, get_usage.

curl -X POST https://verbatimcoach.com/api/v1/generate \
  -H "Authorization: Bearer $VERBATIM_API_KEY" \
  -H "Content-Type: application/json" \
  --max-time 180 \
  -d '{"input": "https://www.youtube.com/watch?v=VIDEO_ID", "formats": ["thread", "carousel"]}'

Response (abridged; formats you did not request are null):

{
  "id": "cm...",
  "object": "generation",
  "status": "completed",
  "source": { "type": "youtube", "url": "https://www.youtube.com/watch?v=VIDEO_ID", "title": "..." },
  "formats": ["thread", "carousel"],
  "outputs": {
    "xThread": ["Post 1", "Post 2", "..."],
    "carousel": { "title": "...", "slides": [{ "kicker": "...", "headline": "...", "body": "..." }], "template": "editorial", "brandColor": "#1F4D3A" },
    "newsletter": null,
    "videoHooks": null
  },
  "links": { "self": ".../api/v1/generations/cm...", "edit": ".../dashboard/generations/cm...", "carouselPdf": ".../api/v1/generations/cm.../pdf" },
  "usage": { "plan": "free", "limit": 3, "used": 1, "remaining": 2, "periodResetsAt": "..." }
}

Point any MCP client that supports Streamable HTTP and custom headers at the endpoint below. Listing tools works without a key; calling them needs one.

{
  "mcpServers": {
    "verbatim": {
      "type": "http",
      "url": "https://verbatimcoach.com/api/mcp",
      "headers": { "Authorization": "Bearer vb_YOUR_KEY" }
    }
  }
}

Every error is JSON: {"error": {"code": "...", "message": "..."}}

• 400 invalid_request: Body is not valid JSON or a field is wrong.
• 401 unauthorized: Missing, invalid or revoked key.
• 402 quota_exceeded: Free monthly drafts used up; upgrade to Pro or wait for the next month.
• 404 not_found: No such draft on this account, or no carousel in it.
• 422 unprocessable_source: The page, video captions or text could not be read or had too little substance.
• 429 rate_limited: Too many requests; retry after the Retry-After seconds.
• 502 upstream_error: The AI writing service failed; try again (not counted).
• 500 internal_error: Unexpected error on our side.

No. The API returns drafts. You or your agent decide what to publish and where.