Built it in someone else's portal? Same answer.

The 'Description' field of your Custom GPT (visible to anyone who opens it).

1. Open your Custom GPT in the GPT Builder.
2. Edit → Configure → "Description".
3. Append on a new line:
      AI Identity: aii://<IDENTITY_ID> · verify at aiidentity.org/p/<IDENTITY_ID>
4. Save. Share your GPT link in /dashboard/identities/<IDENTITY_ID>/website
   so we can fetch it and confirm. Done in ~30 seconds.

If your GPT has Actions, your Action server should also send the
X-AI-Identity header on responses (see /spec/integrations/http).

The Gem's 'Description' field.

1. gemini.google.com → Gems → your Gem → Edit.
2. In "Description", append on a new line:
      AI Identity: aii://<IDENTITY_ID> · verify at aiidentity.org/p/<IDENTITY_ID>
3. Save. Submit the Gem's share URL in your dashboard.

For Gems that call Vertex AI extensions, the extension endpoint should
also serve the X-AI-Identity header on every response.

The Project's 'Custom instructions' field.

1. Open your Project in claude.ai → Settings → "Custom instructions".
2. Append on a new line:
      AI Identity: aii://<IDENTITY_ID> · verify at aiidentity.org/p/<IDENTITY_ID>
3. If the Project is shared, paste the share URL in your dashboard so we
   can verify the reference is live.

Claude Projects don't yet expose a public-share API, so verification
runs via a screenshot/manual review on first registration. Once Anthropic
ships public Projects we'll switch to automated re-verification.

The bot's 'Description' field.

1. poe.com → your bot → Edit bot.
2. In "Description", append:
      AI Identity: aii://<IDENTITY_ID>
3. Save. We confirm via Poe's public bot page.

Project-level CLAUDE.md or .claude/identity.json.

1. In your repo root, add a CLAUDE.md (Claude Code reads this on every session):
      ---
      ai-identity: aii://<IDENTITY_ID>
      passport: <PASSPORT_TOKEN>
      ---
2. Or use a dedicated .claude/identity.json file (preferred for org-wide rollout):
      {
        "ai_identity": "<IDENTITY_ID>",
        "passport": "<PASSPORT_TOKEN>",
        "registry": "https://aiidentity.org"
      }
3. When Claude Code makes outbound HTTP calls (tool calls, MCP, fetch),
   it includes the X-AI-Identity header automatically once configured.

Project-level .cursor/identity.json or .cursorrules.

1. In your repo, add .cursor/identity.json:
      {
        "ai_identity": "<IDENTITY_ID>",
        "passport": "<PASSPORT_TOKEN>"
      }
2. Or append to .cursorrules:
      AI-Identity: aii://<IDENTITY_ID>
3. When Cursor's agent makes outbound calls (Composer / Background Agent /
   tool use), include the X-AI-Identity header on your tool servers so the
   binding is visible to verifiers.

The agent's 'Server URL' (Vapi posts every call event there). Add the header in your handler.

1. In your Vapi dashboard, set the agent's Server URL to your endpoint.
2. In your handler (whatever framework), set the response header on every
   reply so downstream verifiers can confirm the agent's identity:
      X-AI-Identity: <PASSPORT_TOKEN>
3. Also fill the agent's "Bio" field with:
      AI Identity: aii://<IDENTITY_ID>
   so callers who look up the agent in Vapi's public directory see the link.

For inbound phone numbers, also register the surface as wa:+{number} or
tel:+{number} so verifiers can resolve from the phone number alone.

Webhook handler response header + the agent's public name field.

Identical pattern to Vapi:
  1. Webhook responses include X-AI-Identity: <PASSPORT_TOKEN>
  2. Agent's public-facing name/bio includes:
       AI Identity: aii://<IDENTITY_ID>

Agent's 'First message' / 'System prompt' (visible to operators) plus webhook headers.

1. In the ElevenLabs Conversational dashboard, edit the agent's system
   prompt and append:
      Operator's AI Identity: aii://<IDENTITY_ID>
2. If you've configured a webhook for tool calls, set X-AI-Identity on
   every response.

Webhook responses + Zap description.

1. For Zaps that call external APIs, set the X-AI-Identity header on
   any HTTP step's request configuration.
2. Add the AI Identity reference to the Zap's title/description so
   teammates and auditors can see it.

HTTP module headers + scenario description.

1. In each HTTP module's "Headers" config, add:
      X-AI-Identity: <PASSPORT_TOKEN>
2. In the scenario description, note the identity reference for audit:
      AI Identity: aii://<IDENTITY_ID>

HTTP Request node headers + workflow notes.

1. On every HTTP Request node, add:
      X-AI-Identity: <PASSPORT_TOKEN>
   (recommend a credential so it's set once across all workflows)
2. Add the identity reference to the workflow's notes/description.

Agent description + outbound HTTP headers on Action steps.

1. In the Lindy dashboard, edit the agent's description and append:
      AI Identity: aii://<IDENTITY_ID>
2. For Action steps that call external APIs, configure the
   X-AI-Identity header on the request.

Whichever public-facing text field your platform exposes.

Wherever your assistant has a description, bio, system prompt, or
"about" field, paste this on its own line:

   AI Identity: aii://<IDENTITY_ID> · verify at aiidentity.org/p/<IDENTITY_ID>

If your platform also lets you add custom HTTP headers on outbound
calls, set:
   X-AI-Identity: <PASSPORT_TOKEN>

Then submit the public link to your assistant in your dashboard. We'll
do a manual first-time verification and add the platform to this list
if there's enough demand.