StackSwap Open API + MCP

Two integration surfaces. The Open API is the programmatic interface to StackSwap OS — push leads, pull scored queues, tag outcomes, retrain your buyer model from your own stack. The MCP puts seventeen GTM-intelligence tools directly inside Claude Code, Claude Desktop, and ChatGPT — free, no API key, one install command.

Open API

REST · JSON · Bearer token auth. The same enrichment and scoring engine that powers the StackSwap OS UI, accessible programmatically. Push leads in from your CRM, sequencer, or data warehouse; pull the ranked scored queue back into whatever channel your motion runs on. Requires a StackSwap OS subscription (Starter $49/mo · Max $79/mo).

Base URL

https://stackswap.ai/api

Auth: Authorization: Bearer <key> · Keys issued in your StackSwap OS dashboard.

Endpoints

POST/api/leads/enrich

Enrich a batch of companies (up to 100). Pass domain or company name; returns detected tech stack, inferred spend tier, motion type, and ICP fit score against your configured ICP.

{ "companies": [{ "domain": "acme.com" }] }

POST/api/leads/score

Score an already-enriched lead against your ICP. Returns a 0–100 fit score, the top 3 signals driving the score, and the recommended opener angle.

{ "lead": { "domain": "acme.com", "stack": [...] } }

POST/api/model/tag-outcome

Tag a lead's conversion outcome (converted, replied, ghosted, lost). Outcomes queue for the next monthly model retrain — this is how your buyer model gets smarter over time.

{ "domain": "acme.com", "outcome": "converted" }

GET/api/leads/queue

Pull your ranked lead queue — companies scored by fit, sorted highest-first, with opener angles. Paginated. Pipe to your sequencer, CRM, or ad platform.

GET/api/model/status

Returns the current state of your buyer model: last retrain date, accuracy delta vs. baseline, number of tagged outcomes used in training.

Full OpenAPI spec, request/response schemas, and code examples are in the public docs repo. Questions or endpoint requests: nick@stackswap.ai.

MCP — seventeen tools inside your AI client

The StackSwap MCP is a free Model Context Protocol server. One install command puts seventeen GTM-intelligence tools directly inside Claude Code, Claude Desktop, ChatGPT Desktop, or Cursor. Free, no API key, stateless JSON-RPC.

CostFreeTools17AuthNoneTransportHTTP

Install

Claude Code (CLI)

One command in any terminal:

claude mcp add --transport http stackswap https://stackswap.ai/api/mcp

Claude Desktop

Add this block to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows), then restart Claude Desktop:

{ "mcpServers": { "stackswap": { "url": "https://stackswap.ai/api/mcp", "transport": "http" } } }

Cursor / ChatGPT Desktop / other MCP clients

Add a new HTTP MCP server with URL https://stackswap.ai/api/mcp. No authentication. The endpoint responds to GET with a server descriptor for sanity-checking and to POST with JSON-RPC requests. A working TypeScript example client lives in the public docs repo.

https://stackswap.ai/api/mcp

Tools exposed

  • search_tools — Fuzzy-search the StackSwap catalog of ~400 GTM tools by name.
  • find_overlaps — Find curated overlap pairs in a stack with savings estimates.
  • suggest_swaps — AI-native replacement recommendations for legacy tools.
  • scan_stack — Preview StackScan — monthly/annual recoverable spend + top 5 opportunities.
  • recommend_partner — Best-fit affiliate partner for a given GTM need.
  • compare_tools — Head-to-head comparison of two tools.
  • recommend_stack — Greenfield reference stack for an industry vertical.
  • get_tool_details — Full profile for a single tool including partner link.
  • get_vendor_fact_sheet — Full GTM Decision Schema doc — pricing tiers, gotchas, AI capabilities, affiliate terms.
  • compare_tools_n_way — Side-by-side matrix of 2–6 tools in one shot.
  • search_content — Full-text search across the StackSwap operator KB (~50 articles).
  • get_kb_article — Fetch a KB article body as markdown by slug.
  • get_category_landscape — Full map of one GTM category — leaders, runner-ups, skip list.
  • detect_stack_from_text — Infer a GTM stack from a careers page, JD, or site HTML.
  • get_buyer_questions — Operator-authored questions to ask before signing — per-vendor gotchas.
  • get_renewal_strategy — Renewal-negotiation playbook with leverage points and walkaway script.
  • submit_correction — Queue a pricing, feature, or gotcha correction for admin review.

Endpoint: POST https://stackswap.ai/api/mcp · Server descriptor: GET https://stackswap.ai/api/mcp
Protocol versions: 2025-06-18, 2025-03-26, 2024-11-05
Source: docs + schemas + example client at StonesofCreation/stackswap-mcp (MIT). MCP server implementation is closed-source.

FAQ

The StackSwap Open API is a REST JSON API that lets you push leads into StackSwap OS for enrichment and scoring, pull back your ranked lead queue, and tag conversion outcomes to retrain your buyer model — all from your own code, CRM, sequencer, or data warehouse. It is the programmatic surface for StackSwap OS: the same enrichment and scoring engine that powers the hosted UI, accessible via HTTP.

The StackSwap MCP is a Model Context Protocol server that exposes seventeen tools from stackswap.ai directly inside Claude Code, Claude Desktop, ChatGPT Desktop, and Cursor. It covers catalog search, stack overlap detection, AI-native swap suggestions, vendor comparisons, operator KB search, buyer-diligence questions, and renewal playbooks. Free, no API key, stateless JSON-RPC over HTTP.

Yes. The Open API is the programmatic interface to StackSwap OS — enrichment, scoring, and model training all require a subscription (Starter $49/mo or Max $79/mo). The MCP is free and requires no account.

Bearer token authentication via API key. Keys are issued per account in the StackSwap OS dashboard. Include the key as Authorization: Bearer <key> in every request. Rate limits apply per plan tier.

Every time you tag a lead's outcome (converted, replied, ghosted, lost), the signal is queued for the next monthly model retrain. The model learns which stack profiles, spend tiers, and ICP signals correlate with conversion in your specific motion — not the category average. After 3 months of tagging, the model is materially more accurate. After 6, the gap vs. a generic enrichment tool is substantial.

Yes. The MCP endpoint speaks standard JSON-RPC 2.0 over HTTP. You can call it directly via curl or from any code. GET https://stackswap.ai/api/mcp returns the server descriptor; POST with a JSON-RPC envelope invokes any tool.