Funkel API Verified

{
  "service_type": "platform",
  "base_url": "https://funkel.ai",
  "auth_method": "oauth2",
  "auth_config": {
    "rest": {
      "type": "bearer_token",
      "header": "Authorization: Bearer <FUNKEL_PERSONAL_ACCESS_TOKEN>",
      "token_creation": "Settings > Security in Funkel AI dashboard"
    },
    "mcp": {
      "type": "oauth2",
      "flow": "authorization_code_pkce",
      "endpoint": "https://api.funkel.ai/mcp",
      "transport": "streamable_http"
    },
    "notes": "MCP initialize and tools/list permit unauthenticated discovery; tools/call requires auth. REST requires a personal access token; 401 without one is expected."
  },
  "endpoints": [
    {
      "method": "GET",
      "path": "/api/apps",
      "description": "List owned products (read-only; requires bearer token)",
      "base_url": "https://api.funkel.ai"
    }
  ],
  "pricing_model": {
    "type": "subscription",
    "details": {
      "currency": "USD",
      "starting_at": "$99/month",
      "team_tier": "$69/seat/month billed annually (3-19 seats)",
      "enterprise": "20+ seats, custom pricing",
      "free_tier": "not listed"
    }
  },
  "rate_limits": {},
  "capabilities": [
    "Lead discovery from public buying signals",
    "Buyer-fit scoring with retained source evidence",
    "Company-to-person search and CSV company/people imports",
    "LinkedIn, X, and email outreach workflows",
    "Campaign analytics and shared inbox",
    "REST API with OpenAPI documentation",
    "Remote MCP server (Streamable HTTP) supporting Claude and Cursor",
    "OAuth 2.0 with PKCE for MCP clients",
    "Personal access token authentication for REST",
    "Streaming support",
    "Integrations with Clay and CSV imports"
  ],
  "raw_analysis": "Funkel AI is a B2B lead-generation and outreach platform targeting B2B SaaS founders, sales and growth teams, lead-gen agencies, revops, and developers building sales agents. It combines lead discovery from public buying signals, buyer-fit scoring that retains source evidence, company-to-person enrichment, CSV imports, and multi-channel outreach across LinkedIn, X, and email, plus campaign analytics and a shared inbox. Unlike many sales tools, it explicitly exposes a developer surface: a documented REST API (base https://api.funkel.ai, OpenAPI docs at https://funkel.ai/doc/api) using Bearer personal access tokens created in Settings > Security, and a remote MCP server over Streamable HTTP at https://api.funkel.ai/mcp with OAuth 2.0 + PKCE, aimed at MCP clients such as Claude and Cursor. The vendor's benchmark guide indicates REST account reads require a token (401 expected without), while MCP initialize and tools/list allow unauthenticated discovery and tools/call requires authentication. Only one endpoint was surfaced in this submission (GET /api/apps to list owned products, read-only), suggesting the broader API catalog is documented rather than enumerated here. Pricing is subscription-based starting at $99/month, with a Team tier at $69/seat/month billed annually (3–19 seats) and Enterprise at 20+ seats. No free tier was advertised. Maturity looks early-stage: the benchmark entry is auto-registered and vendor-submitted with no test credentials, and only a single endpoint is listed, so API breadth is unverified. Direct alternatives named by the vendor are Clay, Apollo, and Gojiberry. The platform displays developer-oriented traits (OpenAPI docs, OAuth/PAT, MCP with streaming) that distinguish it from typical SDR tooling, and it supports streaming responses. Rate limits were not published. Positioning is a fit-to-score, evidence-backed lead engine that developers can wire into agent stacks via MCP rather than a purely human-facing sales dashboard."
}

1/3 tests passed

{
  "overall": 64,
  "dimensions": {
    "token_efficiency": 8.0,
    "first_try_success": 6.0,
    "response_parseability": 8.5,
    "error_clarity": 6.0,
    "doc_quality": 5.0,
    "auth_simplicity": 7.5,
    "latency": 6.5,
    "consistency": 5.5
  },
  "pricing_normalized": {
    "currency": "USD",
    "starting_monthly": 99,
    "annual_seat_monthly": 69,
    "free_tier": false,
    "enterprise_custom": true
  },
  "issues": [
    "No free tier listed — agents cannot recommend a low-risk trial path; users must commit at $99/mo minimum",
    "Website root returned 404 during probe — landing page may be hosted elsewhere or route changed; hurts first-try trust",
    "No llms.txt published — agents lack a canonical, token-efficient entry point for the platform",
    "robots.txt content-signal restrictions may limit agent-side content retrieval, adding friction to doc discovery",
    "Latency ~500-720ms on probed endpoints — acceptable but not snappy; sits in the mid band",
    "Documentation depth not confirmed via probe — relies on third-party knowledge for onboarding edge cases"
  ],
  "recommendations": [
    "Publish an llms.txt summarizing capabilities, auth flows (OAuth PKCE for MCP, PAT for REST), pricing tiers, and quickstart links to reduce agent token spend and improve first-try setup",
    "Fix the 404 on the primary website route (or ensure the canonical URL agents are given resolves 200) to preserve first-try success",
    "Introduce a free trial or sandbox tier (e.g., 14-day or limited-credit) so agents can recommend an evaluation path without pricing commitment",
    "Expose a status page and uptime SLA note in docs to bolster consistency perception, especially for MCP/API consumers",
    "Provide structured error examples (HTTP codes, JSON error schema, retry guidance) in the OpenAPI spec to raise error clarity",
    "Add self-serve signup with magic-link/SSO and clear OAuth scopes documentation to keep auth simplicity high",
    "Clarify robots.txt content-signal policy for AI agents reading public docs, or host an agent-friendly docs mirror"
  ]
}