Vouch MCP server
Connect your AI IDE (Cursor, Claude Desktop, Windsurf, Zed) to your Vouch workspace over the Model Context Protocol. Your interviews, assumptions, and segments become first-class context for any AI assistant you use, and the assistant can write back — log an interview, add an assumption, save a prospect.
Endpoint:
https://vouch.specky.space/api/public/mcpTransport: MCP Streamable HTTP (POST)
Auth:
Authorization: Bearer vch_…Why connect Vouch over MCP
Your customer interviews, ranked assumptions and segment scores are the highest-signal context for any product decision an AI assistant can help with. Over MCP you can:
- Ground roadmap or spec drafts in real customer quotes (every excerpt is citable).
- Ask "what do we already know about X?" before scheduling another interview.
- Cross-check a proposal against the project's open assumptions.
Setup
- Sign in at vouch.specky.space.
- Open /settings/mcp, click Generate token, and copy the
vch_…value (shown once). - Add the server to your IDE config below.
Cursor / Claude Desktop / Windsurf
{
"mcpServers": {
"vouch": {
"url": "https://vouch.specky.space/api/public/mcp",
"headers": {
"Authorization": "Bearer vch_YOUR_TOKEN_HERE"
}
}
}
}Zed
{
"context_servers": {
"vouch": {
"command": {
"path": "npx",
"args": [
"-y", "mcp-remote",
"https://vouch.specky.space/api/public/mcp",
"--header", "Authorization: Bearer vch_YOUR_TOKEN_HERE"
]
}
}
}
}Tools exposed
| Tool | Purpose |
|---|---|
validate_idea | Turn an idea + target customer into 3–5 ranked FeDeVi assumptions to test. |
get_methodology | The Vouch validation method, so an assistant answers 'how should I validate?' correctly. |
list_industry_templates | Industry-specific starting points and the assumptions they come with. |
browse_public_gallery | Real public ideas other founders have validated, as examples. |
start_validation | Hand the founder into Vouch with their idea pre-filled. |
get_share_kit | Ready-to-post LinkedIn / X / newsletter copy about what they learned. |
list_my_projects | Your ideas with id, name, idea statement and target customer. |
get_my_project | One project in full — assumptions, interviews, segments, pilots and LOIs. |
create_project | Start a new idea in your workspace. |
add_assumptions | Append risky assumptions with FeDeVi category and importance. |
log_interview | Save a customer interview (name, role, notes or transcript) to a project. |
list_prospects | Outreach prospects saved for a project. |
add_prospects | Save people to contact — e.g. from LinkedIn — as prospects. |
update_prospect_status | Mark a prospect contacted, replied, booked or not interested. |
draft_social_post | Save a LinkedIn or dev.to draft. Drafts are never auto-published. |
list_social_posts | Your drafts and published posts with their stats. |
list_experiment_methods | Experiment methods (fake door, smoke test, concierge, mock sales, LOI, A/B…) with templates. |
run_validation_sprint | Design an honest experiment for up to 6 of your ideas at once and save each plan. |
create_test_page | Draft the waitlist test page for a sprint idea (private until published). |
publish_test_page | Make a sprint test page public so you can share it. |
get_sprint_results | Signups per test page against your success and stop thresholds. |
get_gtm_plan | Your 5 go-to-market steps, what's done and what's next. |
update_gtm_step | Tick a go-to-market step done or undone. |
All tools are scoped to the authenticated user — tokens never see another user's data.
Recommended prompts
- "Use Vouch to list my projects, then summarize what's been validated for the [name] idea."
- "Search interviews in project [id] for 'pricing' and quote the three strongest objections."
- "Read assumptions for project [id]. Which Desirable assumptions still have weak evidence?"
- "Get segments for project [id] and recommend which one to pursue first, citing scores."
Security
- Tokens are SHA-256 hashed at rest; the raw
vch_…value is only displayed once. - Revoke anytime at /settings/mcp. Revoked or invalid tokens return HTTP 401 with a
WWW-Authenticatehint. - Suspicious activity (high failure rate, multiple IPs per token, burst usage, use after revocation) is surfaced in /settings/mcp and the admin console.
- Write tools act as you, never as someone else: every project, interview, prospect and draft a tool touches is checked against your own account. Social posts are saved as drafts — Vouch never publishes for you.
- Use is tracked: each token shows how many times it was used and when, you can revoke it in one click, and unusual patterns (many failed checks, several IPs, usage bursts, use after revocation) are surfaced in /settings/mcp and the admin console.
Error reference
| Code | Meaning | Fix |
|---|---|---|
401 missing | No Authorization header. | Add Authorization: Bearer vch_… from /settings/mcp. |
401 invalid_token | Token not recognized. | Generate a new one at /settings/mcp. |
401 token_revoked | Token was revoked. | Generate a new one and update your IDE config. |
405 | GET or DELETE attempted. | MCP Streamable HTTP uses POST. |
406 | Client missing Accept header. | Use a real MCP client; don't hand-roll fetch. |
Spec & implementation
- MCP Streamable HTTP spec: modelcontextprotocol.io
- Built with mcp-tanstack-start
- Machine-readable companion: /mcp.md