MCP SERVER · LIVE · /api/mcp

Give your AI agent
permission to act.

AI agents are starting to fill forms and share records for families, and today they do it by screen-scraping credentials and guessing at what was allowed. When a college, a scholarship program, or an insurance agent asks what was actually authorized, nobody can prove it. Eformogi is the records office that fixes this: agents get scoped, policy-governed access to your verified profile and consent vault, every action is logged, and permission becomes something you can prove.

Read the MCP docs

Endpoint: https://eformogi.com/api/mcp · Auth: Bearer efg_…

AI is about to act for you. The question is who governs it.

An AI agent that fills a form on your child's behalf, sends a transcript to a college, submits a scholarship application, or answers an insurance agent's request for records is taking a real action with real consequences. Without a records office, every agent ends up screen-scraping credentials and guessing at consent. Eformogi gives agents a clean, signed, policy-aware interface, and it gives you a single place to grant, scope, and revoke their access.

How it works

1
Generate a personal API key
In Settings, you create a scoped key — choose which categories the agent can read or write (medical, education, finance, contact, identity).
2
Connect your agent
Drop the key into Claude, ChatGPT, Gemini, or any MCP-compatible client. The agent now sees your vault — only what you allowed.
3
Set your policy
For each category, decide: auto-approve, ask first, or never. The policy engine enforces it for every request the agent makes.
4
Review the activity feed
Every read, write, and share the agent makes is logged — with timestamp, scope, and outcome. Revoke access anytime.

The MCP toolset

Six tools. Fully audited. No surprises.

read_profile()
Return scoped profile fields the user has approved for this agent.
list_consents()
Show every active consent the user has granted, with expiry and scope.
create_consent()
Request new consent on the user's behalf — they approve in-app, signed and logged.
check_policy()
Ask whether an action is allowed before doing it. Auto, ask, or never — per category.
log_action()
Write to the agent activity feed. Every read, write, and share is auditable.
search_orgs()
Discover trusted organizations the user can share verified data with.

Three lines to first call

JSON-RPC 2.0 over POST. Standard MCP — no SDK required.

curl -X POST https://eformogi.com/api/mcp \
  -H "Authorization: Bearer efg_..." \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'

Full protocol spec at /docs/mcp · OpenAPI at /.well-known/openapi.json

Built on consent. Not credentials.

Other "agent integrations" hand AI your password and hope. Eformogi gives agents a scoped key tied to a policy you control. Revoke it in one click. See every action it took. The records office is yours, not the model's.