Brightbean - Youtube intelligence MCP for AI agents

Created By
BrightBean6 days ago
Youtube intelligence for AI agents - score titles & thumbnails, surface niche content gaps, benchmark channels and videos
Overview

BrightBean - YouTube intelligence MCP server

Homepage: https://brightbean.xyz Docs: https://brightbean.xyz/docs/mcp/ MCP endpoint: https://api.brightbean.xyz/mcp Transport: Streamable HTTP Auth: Bearer API key (bb_…) or OAuth 2.1 (Claude Desktop native connector) Pricing: Free tier - 350 credits/month, no card required. Paid plans from $19/mo (Hobby) → $399/mo (Growth). Hosting: Fully hosted by BrightBean, no install, no local dependencies.

BrightBean is a YouTube intelligence API exposed as an MCP server. It predicts click-through rate for title and thumbnail combinations, scores video hooks (the first ~6 seconds), surfaces ranked content gaps inside YouTube niches, and benchmarks channels and individual videos against their niche baselines. Calls draw from a shared credit balance the agent sees inline (credits_remaining is returned with every successful call) so agents can self-regulate spend.

Tools (6)

ToolCreditsWhat it does
score_packaging1–3Score a YouTube title and/or thumbnail for predicted CTR. Mode auto-detected: title only = 1cr, thumbnail only = 2cr, combined = 3cr.
score_video_hook10Analyze the first ~6 seconds of a YouTube video. Returns hook archetype + 5 dimension scores (clarity, specificity, tension, visual_energy, pace) + suggestions.
research_content_gaps5Ranked content opportunities for a YouTube niche. Filter by gap_type (underserved / stale / competitive), min_score, limit.
list_niches1List all catalogued YouTube niches with gap counts. Returned slugs are valid inputs to research_content_gaps.
benchmark_channel5Benchmark a YouTube channel against its niche. Fetches the last 10 videos, returns engagement percentiles, title patterns, exemplar channels.
benchmark_video3Benchmark a single YouTube video against its niche. Engagement percentiles + title-pattern fit.

Credits are charged only on success (2xx responses). Failures cost nothing.

Connect in 60 seconds

Step 1 - Get an API key. Sign up at https://app.brightbean.xyz, open https://app.brightbean.xyz/dashboard/keys/, click Create key. The key starts with bb_ and is shown once.

Step 2 - Pick a client.

Claude Desktop / Claude Code (config file)

Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

```json { "mcpServers": { "brightbean": { "command": "npx", "args": [ "mcp-remote", "https://api.brightbean.xyz/mcp", "--header", "Authorization:${BRIGHTBEAN_KEY}" ], "env": { "BRIGHTBEAN_KEY": "Bearer bb_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx" } } } } ```

Fully quit Claude Desktop (⌘Q) and reopen. The six BrightBean tools appear under the MCP tools control next to the message box.

The key goes in env (not inline in --header) because mcp-remote mis-parses a --header value containing a space.

Cursor

Edit ~/.cursor/mcp.json (or .cursor/mcp.json for project scope):

```json { "mcpServers": { "brightbean": { "url": "https://api.brightbean.xyz/mcp", "headers": { "Authorization": "Bearer bb_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx" } } } } ```

Save - Cursor connects and lists the tools.

Claude Desktop - Native custom connector (no config file)

If you'd rather sign in through a browser instead of pasting a key:

  1. Customize in the left sidebar → Add custom connector.
  2. Name: BrightBean. Server URL: https://api.brightbean.xyz/mcp. Click Add.
  3. A BrightBean sign-in window opens - log in / sign up → approve access. Tool calls draw credits from the account you signed in with.

Generic MCP client (LangChain, agent runtimes, custom)

  • Endpoint: https://api.brightbean.xyz/mcp
  • Headers: Authorization: Bearer bb_…
  • Transport: Streamable HTTP

Verifying

Ask your client:

Use the brightbean tools to score the title "I bought a $0.99 keyboard so you don't have to" with thumbnail https://i.ytimg.com/vi/EXAMPLE/hqdefault.jpg in the tech-reviews niche.

A working setup returns a score envelope with score_id and credits_remaining populated.

Tags

youtube, analytics, ctr, thumbnails, video, creator-tools, content-research, benchmarks, remote, hosted, oauth

Category

Analytics

Notes for the reviewer

  • BrightBean is a hosted remote MCP there is no install command and no public source repository at this time. The endpoint above is the canonical surface.
  • CORS is intentionally disabled on /mcp; MCP clients are expected to be server-side processes or local desktop apps.
  • Full tool reference: https://brightbean.xyz/docs/mcp/

Server Config

{
  "mcpServers": {
    "brightbean": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://api.brightbean.xyz/mcp",
        "--header",
        "Authorization:${BRIGHTBEAN_KEY}"
      ],
      "env": {
        "BRIGHTBEAN_KEY": "Bearer bb_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
      }
    }
  }
}
Project Info
Created At
6 days ago
Updated At
6 days ago
Author Name
BrightBean
Star
-
Language
-
License
-
Category

Recommend Servers

View All
AI Work Market — USDC settlement rails for AI labor on Base Mainnet)
@Dario (DME)

AI Work Market is a USDC escrow protocol on Base Mainnet, designed for autonomous AI agents to find work, post jobs, and settle payments without humans in the loop. This MCP server exposes 10 tools: **Escrow lifecycle** - `create_intent_quote` — get calldata + gas estimate for funding a new escrow intent - `submit_proof_quote` — get calldata for the seller to submit a proof URI - `release_funds_quote` — get calldata for the buyer to release payment (or claim/refund) **x402 single-call binding** - `x402_consume` — replaces the 5-step x402 flow with one HMAC-signed POST that returns a delivery URL **Onboarding & discovery** - `agent_onboard` — generate a signed agent card with marketplace attestation - `agent_search` — tf-idf search over the live agent catalog - `agent_reputation` — server-side reputation from on-chain Released/Refunded/Disputed events **Live state** - `system_status` — live on-chain state (nextIntentId, accumulatedFees, contract balance, owner) - `escrow_rules` — contract semantics, lifecycle, call guides, failure modes - `events_subscribe` — SSE stream of new on-chain intent events All endpoints are serverless (Vercel) and return their schema on GET. No browser, no wallet UI required for an agent to integrate. The protocol takes a 1% commission on every settlement; the rest goes to the seller. The full AgentCard is at `/.well-known/agent-card.json` (A2A-compatible). The OpenAPI 3.0.3 spec is at `/.well-known/openapi.json` with `components.securitySchemes` (none, hmacX402). `robots.txt` allows GPTBot, ClaudeBot, anthropic-ai, PerplexityBot, Google-Extended, Applebot-Extended, CCBot, Amazonbot.

7 hours ago
Bring your real authenticated browser session to AI coding agents. Local-first MCP server + Chrome MV3 extension. No cloud. No telemetry.
@Cubenest

peek records the user's actual logged-in browser (DOM via rrweb, console events, network metadata, optional response bodies via opt-in Deep capture) through a Chrome MV3 extension. The extension ships events through a native-messaging stdio bridge to a local MCP server (peek-mcp), which persists them to a SQLite database at ~/.peek/sessions.db. AI coding agents (Claude Code, Cursor, Cline, Windsurf) read sessions from the database via 10 MCP tools: Tool What it does list_recent_sessions List recently recorded sessions (id, origin, ts, event count). get_session_summary LLM-readable narrative summary of a session. get_session_console_errors Console errors recorded in a session. get_session_network_errors Failed/notable network requests in a session. get_user_action_before_error Last N user actions before a console error. generate_playwright_repro Generate a runnable Playwright test from a session. get_dom_snapshot Reconstruct the DOM at a given timestamp. query_dom_history Timeline of attribute/text changes for a selector. request_authorization Side-panel consent for write actions (Level 3). execute_action Dispatch a UI action (gated by permission level + destructive blocklist). Why local-first matters Every other "browser session for AI" tool ships to a vendor cloud. peek's SQLite + extension live on the user's machine — no remote endpoints, no telemetry. The privacy policy (docs/peek/PRIVACY_POLICY.md) is the source of truth. Install # 1. Add the MCP server to Claude Code claude mcp add peek -- npx -y @peekdev/mcp # 2. Install the Chrome extension from the Chrome Web Store # (link added once the CWS listing is approved)

a day ago