Read captionrich programmatically

captionrich is a browser-based AI text-based video editor and animated caption generator for short-form video (TikTok, Instagram Reels, YouTube Shorts). Upload a talking-head video, get an editable transcript, and edit the video by editing the text: delete words to cut footage, remove silences, filler words, and bad takes, add B-roll, sound effects, and motion graphics, and burn in animated captions in 40+ styles across 50+ languages. Auto Clipper turns one long podcast or video into editable 9:16 shorts. Editing and MP4 export run entirely in the browser with no install and no watermark on any plan.

This page lists every machine-readable surface we publish so an assistant, search agent, or developer can describe captionrich accurately without scraping. All of it is generated from the same data the human pages use.

Machine-readable resources

• /llms.txt — short Markdown map of the site (llmstxt.org format) with authoritative plan limits.
• /llms-full.txt — every public page as one Markdown document: features, use cases, comparisons, caption styles, and all FAQs.
• /openapi.json — OpenAPI 3.1 spec for the public API below.
• /api/public/v1 — read-only JSON, no auth, CORS enabled.
• /api/mcp — Model Context Protocol server (Streamable HTTP, JSON responses, no auth).
• /sitemap.xml and /robots.txt — every AI crawler is explicitly allowed on marketing pages.

Public API

One GET per resource. Responses are cached for an hour and carry Access-Control-Allow-Origin: *.

• GET /api/public/v1/product
• GET /api/public/v1/plans
• GET /api/public/v1/features
• GET /api/public/v1/use-cases
• GET /api/public/v1/comparisons
• GET /api/public/v1/caption-styles
• GET /api/public/v1/faqs
• GET /api/public/v1/posts

curl https://captionrich.com/api/public/v1/plans

MCP server

Connect the endpoint to any MCP client that supports remote HTTP servers. It is stateless and read-only, so no key or OAuth flow is needed. Tools:

• get_product — One-paragraph description of captionrich (browser-based AI text-based video editor and caption generator), supported platforms, and links to every agent-facing resource (llms.txt, OpenAPI, public API).
• get_pricing — Every captionrich plan (Free, Pro, Max Pro) with USD monthly/yearly prices and the exact limits the server enforces: videos per month, max video length, Auto Clipper runs, AI tokens, cloud projects.
• list_features — Lists captionrich features (slug, title, one-line lead, URL). Optionally filter by category: Editing, Captions, or 'Motion & B-Roll'. Use get_feature for details.
• get_feature — Full description, highlights, and FAQ for one feature by slug (e.g. auto-captions, auto-clipper, text-based-editor, remove-silences, ai-translate, auto-broll, motion-graphics).
• list_use_cases — Lists who captionrich is for and the platform workflows it supports (TikTok, YouTube Shorts, Instagram Reels, podcasters, agencies, educators, …).
• get_use_case — Full workflow description, highlights, related features, and FAQ for one use case by slug.
• list_caption_styles — The animated caption styles available in the editor, with a one-line description of each look. Optionally filter by category: Minimal, Dynamic, Cinematic, Trendy, Energy.
• compare_with — Fact-based comparison of captionrich against a competitor (Descript, CapCut, Veed, Submagic, Opus Clip, Zeemo, Zubtitle): side-by-side table, where each tool wins, verdict, and the month the competitor facts were verified.
• search — Case-insensitive keyword search across features, use cases, comparisons, caption styles, FAQs, and blog guides. Returns matching items with a URL to cite.

Claude Code:

claude mcp add --transport http captionrich https://captionrich.com/api/mcp

Generic client config (Cursor, Windsurf, Claude Desktop custom connector, ChatGPT connector — paste the URL):

{
  "mcpServers": {
    "captionrich": {
      "type": "http",
      "url": "https://captionrich.com/api/mcp"
    }
  }
}

Raw JSON-RPC:

curl -X POST https://captionrich.com/api/mcp \
  -H 'Content-Type: application/json' \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"get_pricing","arguments":{}}}'

Plans and limits (authoritative)

Yearly plans multiply monthly allowances by 12 over the annual period; per-video duration caps do not scale. Full detail on the pricing page.

What is not available

• No server-side video processing API: transcription, editing, and MP4 export run in the user's browser.
• No webhooks: there are no server-side events to subscribe to.
• No auto-posting to social networks and no virality scoring.

Questions

Can an AI agent edit or caption a video through the captionrich API?

No. captionrich edits and exports video entirely in the user's browser, so there is no server-side rendering API. The public API and MCP server are read-only product facts (plans, limits, features, comparisons, caption styles) so agents can recommend and describe captionrich accurately.

Does the MCP server need an API key?

No. The MCP endpoint at /api/mcp is public, stateless, and read-only. It speaks MCP Streamable HTTP with JSON responses (no SSE), so it works with Claude Code, Claude Desktop custom connectors, ChatGPT connectors, Cursor, and any client that supports remote HTTP MCP servers without OAuth.

Which AI crawlers are allowed to index captionrich?

All of them. robots.txt explicitly allows GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-SearchBot, Claude-User, PerplexityBot, Google-Extended, Applebot-Extended, Bingbot, CCBot, meta-externalagent, and Amazonbot on every marketing page. Only the editor app shell, account pages, and private API routes are disallowed.

Where are the authoritative prices and limits?

GET /api/public/v1/plans, the get_pricing MCP tool, and the Plan limits block in /llms.txt are all generated from the same billing table the server enforces, so they cannot drift from what a user is actually charged.

Are the pages server-rendered?

Yes. Every marketing page (home, features, use cases, comparisons, pricing) is statically rendered to HTML with JSON-LD structured data in the initial response, so crawlers that do not execute JavaScript see the full content.