Home

IGL, Identity Governed Logic

IGL makes every AI action answer four questions before it runs: who is acting, what are they allowed to do, why, and what exactly happened. The runtime refuses anything that is missing one of those. The record it keeps is signed and hash chained, so nobody can quietly change it afterwards.

This wiki is the front door. It is written for two readers at once: the person who will build on IGL today, and the person who has to decide whether their company should trust an AI system with real work. Each page starts with the plain version and then goes under the hood.

Try it in sixty seconds

No account, no sales call. The sandbox issues you a tenant and a key on the spot.

curl -s -X POST https://sandbox.id-me.ai/sandbox/signup \
  -H "content-type: application/json" \
  -d '{"organization":"My Company","email":"you@example.com"}'

Then ask the platform who you are, under a purpose (put the apiKey and tenantId from the answer in IGL_KEY and IGL_TENANT first):

curl -s "https://sandbox.id-me.ai/v1/identities/me?purpose=identity-introduction" \
  -H "authorization: Bearer $IGL_KEY" -H "x-igl-tenant-id: $IGL_TENANT"

The same two steps in Windows PowerShell, where curl is an alias for Invoke-WebRequest and the bash lines above do not run as written:

$r = Invoke-RestMethod -Method Post -Uri "https://sandbox.id-me.ai/sandbox/signup" `
  -ContentType "application/json" -Body '{"organization":"My Company","email":"you@example.com"}'
$env:IGL_KEY = $r.apiKey; $env:IGL_TENANT = $r.tenantId
Invoke-RestMethod -Uri "https://sandbox.id-me.ai/v1/identities/me?purpose=identity-introduction" `
  -Headers @{ authorization = "Bearer $env:IGL_KEY"; "x-igl-tenant-id" = $env:IGL_TENANT } | ConvertTo-Json -Depth 8

What comes back is not just data. It is the nodes you were allowed to see, a count of what was withheld and why, and a signed governance receipt you can verify with the public key at auth.id-me.ai without trusting the service. The full walk-through is in Quickstart: the API.

What you get

The four guarantees

If you are	Start here	Then
A developer	Quickstart: the API	Purposes and Projections, MCP Integration
Running AI assistants in a company	Why IGL	Use Cases, Operator Console
Responsible for security or compliance	Security Model	Governance Receipts and Audit
Curious about the language itself	Quickstart: the Runtime	Core Concepts, Glossary

What is live right now

Surface	Address	What it is
Site and documentation	id-me.ai	The product site, generated API reference, interactive User Center and developer console
Production API	`https://api.id-me.ai/v1`	51 operations under the public contract, authenticated, `igl_live_` keys
Sandbox	`https://sandbox.id-me.ai/v1`	Same code, separate database, self-serve signup, `igl_test_` keys
Token endpoint and keys	`https://auth.id-me.ai`	OAuth 2.0 client credentials and the JWKS used to verify receipts
Source	github.com/SMEPro-Technologies-LLC/igl	Apache 2.0. The runtime, the contract, the service, the site, the console

Every one of those is deployed from the repository by GitHub Actions after its tests pass; the API deploy runs all 51 operations against a local build before the sandbox is touched, and smoke tests the sandbox before production. Nothing is edited by hand in production.

Why people share this

Most AI governance is a document. IGL is syntax. A program that does not declare its identity does not parse. An action outside its boundary does not run; it is not downgraded to a warning. A result that cannot cite the records it came from is refused with a 422 instead of being returned. And every answer ships with a receipt that a third party can check in a few lines of code.

That is the whole pitch: the referee is built into the game. If you want the longer argument, read Why IGL and the founder's essay, Identity Governed Logic.

Pages

Getting started: Quickstart: the API, Quickstart: the Runtime, MCP Integration

Understanding IGL: Why IGL, Core Concepts, Ricorsa and IGL, Purposes and Projections, Governance Receipts and Audit, Policies, Agents and Workflows

Running it: Operator Console, Deployment and Operations, Security Model

Reference: Use Cases, FAQ, Glossary, Roadmap and Contributing, Brand Assets

IGL, Identity Governed Logic. Code and contract under Apache 2.0; IGL and the logo are trademarks of SMEPro Technologies LLC. Platform at id-me.ai. Security reports: security@id-me.ai. This wiki describes what runs today; see Roadmap and Contributing for what is planned.