Captapi

Captapi is one API for structured data from YouTube, TikTok, Instagram, and Facebook. One key works across all platforms. No OAuth, no per-platform SDKs. Responses are clean JSON and cached for 24h (repeat calls cost 0 credits). 62 endpoints total.

• Base URL: https://api.captapi.com
• Docs: https://captapi.com/docs · Full machine reference: https://captapi.com/llms-full.txt

Step 0 — Get the API key (ask the human)

Using Captapi requires a capt_live_... API key, and creating one requires a human (sign-up cannot be automated). Before doing anything else:

1. If you do not already have a key, ask the user: "Create a Captapi API key at https://captapi.com/dashboard/api-keys (100 free credits, no card) and paste it here."
2. Never guess, invent, or try to sign up for a key. Store the key the user gives you and use it for all requests.

How to call Captapi

Every endpoint is a single authenticated GET request. Pass parameters as URL query params (URL-encode values). Send the key as a Bearer token:

curl "https://api.captapi.com/v1/youtube/transcript?url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DdQw4w9WgXcQ" \
  -H "Authorization: Bearer capt_live_..."

Response shape:

{ "success": true, "cached": false, "creditsUsed": 2, "data": { "language": "en", "text": "..." } }

If you have the MCP server connected instead, call the tool named in the tables below (e.g. youtube_transcript) with the same parameters — no URL building needed.

Use from the command line (@captapi/cli)

For shell tasks, scripts, or when no MCP client is available, the official CLI calls the same API. Every endpoint is a subcommand (the tool name with dashes), parameters are flags, and results print as JSON to stdout:

npx @captapi/cli login                 # save the human-provided key to ~/.captapi/config.json
npx @captapi/cli balance               # remaining credits
npx @captapi/cli list                  # all 62 commands
npx @captapi/cli youtube-transcript --url "https://www.youtube.com/watch?v=dQw4w9WgXcQ"
npx @captapi/cli tiktok-comment-replies --url "<video>" --comment_id "<id>" --limit 20

The CLI reads the key from ~/.captapi/config.json (via login) or the CAPTAPI_API_KEY env var. npx @captapi/cli agent add cursor writes the MCP config into Cursor/Claude for you. Same auth, credits, and error codes as the REST API.

Use in n8n workflows (n8n-nodes-captapi)

For no-code/low-code automations, the official n8n-nodes-captapi community node exposes all 62 endpoints in n8n. Install it from Settings → Community Nodes (package n8n-nodes-captapi; self-hosted: npm install n8n-nodes-captapi, then restart). Create a Captapi API credential with the human-provided capt_live_... key, add the Captapi node, pick a Platform and Operation, and it returns the same structured JSON as the REST API for downstream nodes.

Use in Make.com scenarios (custom app)

For Make.com (Integromat), the Captapi custom app exposes all 62 endpoints as action modules grouped by platform. Create a Captapi API Key connection with the human-provided capt_live_... key (verified against /v1/account/limits), then drop the module you need into a scenario, fill in the url (or search query) and optional limit, and it returns the same structured JSON data as the REST API for downstream modules.

Use on Apify (BYO-key Actor)

On Apify, the Captapi Actor is a bring-your-own-key wrapper around the REST API (no scraping). Set the apiKey input to the human-provided capt_live_... key, choose an operation (any of the 62 endpoints), fill the fields it needs (url / search query / limit / ...), and the Actor returns one dataset item with the same structured JSON data as the REST API. Credits are billed to the user's own Captapi account. The Actor is also callable through Apify's MCP server (mcp.apify.com), so agents already connected to Apify can run it by name.

Choosing the right endpoint

• Single piece of content (one video / reel / post): use *_transcript, *_summarize, *_video_details / *_details, or *_comments with the content url.
• A creator / account: use *_channel_details (stats) or *_channel_posts / *_channel_reels (their content) with the profile url.
• Discovery: use *_search, *_hashtag_search, or *_user_search with a q query.
• Downloads: use *_video_download for a direct media URL.

Parameter rules (important gotchas)

• url — pass the full public URL of the video/reel/post/profile. Each endpoint's table notes the expected URL type.
• q — search query or keyword (min 2 chars). For hashtag endpoints, pass the tag without #.
• limit — optional; controls how many items list/search/comment endpoints return. Billed per result, so request only what you need. Defaults and maxes vary per endpoint.
• language — optional ISO code (e.g. en) for the YouTube transcript/summary endpoints (incl. Shorts); defaults to auto-detect. Other platforms' transcript/summary endpoints take only url.
• comment_id — required for *_comment_replies; get it from the corresponding *_comments response.

Credits & errors

• Each endpoint costs a fixed number of credits (see tables). Cached results (within 24h) cost 0. Failed or empty results are never charged.
• Error responses are non-2xx with { "detail": "..." }:
• 401 — missing/invalid key. Re-check the key with the user.
• 402 — out of credits. Tell the user to top up at https://captapi.com/dashboard/billing.
• 422 — unprocessable (e.g. the video has no captions). Not charged; do not retry blindly.
• 429 — rate limited. Back off and retry after a short delay.

Endpoint reference

YouTube

TikTok

Instagram

Facebook

Generated from the Captapi catalog. Do not edit by hand — run node generate.mjs.