Generate Insights API retires on October 30, 2026: move to the Signal API

Effective date: October 30, 2026. On that date, every Generate Insights API version (v1.0.0 to v1.4) returns HTTP 410. The Signal API replaces it. The Generate Content API does not change.

We are retiring the Generate Insights API on October 30, 2026. The Signal API replaces it, and it does far more.

The Generate Insights API gave you a short list of insights for one contact. The Signal API gives you the full Autobound signal graph: 35+ signal types and 900+ subtypes on companies and people. You can enrich one record, search the whole market, track buyer intent, and use it all from your AI assistant through MCP. This page shows what you get and how to move in 5 steps.

What is changing

• On October 30, 2026, we retire every version of the Generate Insights API: v1.0.0, v1.1, v1.2, v1.3 and v1.4.
• From that date, each call to https://api.autobound.ai/api/external/generate-insights/... returns HTTP 410 with error code 4006.
• Until then, the Generate Insights API works as it does today.

Not affected: the Generate Content API (v3.5, v3.6), AI Studio, and the Signal API. You are affected only if your code calls a URL that starts with https://api.autobound.ai/api/external/generate-insights/.

Why the Signal API is a big upgrade

Free discovery endpoints help you build fast: list signal types, list filters, signal type history, credit costs, account and request logs.

Simple pricing. You pay only for results. Enrichment costs 2 credits per returned signal. Zero-result requests are free. See Credits.

We add new signal types and parameters often. Follow the changelog to see each release.

Use it from your AI assistant (MCP)

The Autobound MCP server lets Claude, ChatGPT, Cursor, Claude Code and any MCP client query the Signal API in plain language.

• Server URL: https://mcp.autobound.ai/mcp
• Claude Code, in one line:

claude mcp add --transport http autobound https://mcp.autobound.ai/mcp

• Example prompts: "What signal types do you have?" or "Which fintech companies announced layoffs in the last 30 days?"

Full setup for each client: Autobound MCP.

Core differences to prepare for

The Signal API is a new API, not a new version of Generate Insights. Plan for these 5 changes.

1. A new JSON shape. Your parser must change.

• Results come in signals (company call) or contact_signals (contact call), not insights.
• Each signal has signal_type, signal_subtype, detected_at and a data object.
• The fields inside data depend on the signal type. Each type has its own guide in the signal catalog, and the schema page shows the shared envelope.
• The response also has company, contact, signal_summary and coverage.

2. Insight subtypes become signal types and signal subtypes. The old API used one flat insightSubtype. The Signal API uses two levels: a signal type (for example news) and a signal subtype (for example receives_award). Today there are 35+ signal types and 900+ subtypes. Get the live list from List all signal types. It is free.

Subtype matching ignores case. An unknown subtype returns zero results, not an error. So check your values against the live list.

3. Raw signals, not ready-made insights. The old API returned insight text, talking points, and (in v1.3 and v1.4) a ranking for your seller context. The Signal API returns structured signals with dates and source data. You pick and rank them. Many types include a data.summary. To turn signals into an email, use the Generate Content API.

4. Insight Control settings do not apply. Org settings and custom insights from Insight Control shaped the old API output. The Signal API ignores them. Use signal_types, signal_subtypes, detected_after and limit instead.

5. New pricing and a new key. The old API charged per request. The Signal API charges 2 credits per returned signal, and zero-result requests are free. It also uses its own API key.

How to move (5 steps)

1. Get a Signal API key

Sign in to the Signal API developer portal and create a key. You can also ask your account manager for one.

The Signal API key is separate from your current Autobound API key. Your current key keeps working for the Generate Content API.

2. Change the endpoint

Send the key in the X-API-KEY header, as today. Authorization: Bearer <key> also works.

3. Map the request fields

4. Map the response fields

5. Test before October 30

Run your main use case against the Signal API. Check the result count, the fields you read, and your credit use. The response headers x-credits-charged and x-credits-remaining show the credits for each call. The onboarding guide walks through a first integration.

Example

Before (Generate Insights):

curl -X POST "https://api.autobound.ai/api/external/generate-insights/v1.4" \
  -H "X-API-KEY: $AUTOBOUND_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"contactEmail": "[email protected]", "contactCompanyUrl": "stripe.com"}'

After (Signal API, company signals):

curl -X POST "https://signals.autobound.ai/v1/companies/enrich" \
  -H "X-API-KEY: $SIGNAL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"domain": "stripe.com", "signal_types": ["news", "hiring-trends"], "detected_after": "2026-07-01T00:00:00Z", "limit": 20}'

After (Signal API, contact signals):

curl -X POST "https://signals.autobound.ai/v1/contacts/enrich" \
  -H "X-API-KEY: $SIGNAL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"contact_email": "[email protected]", "limit": 20}'

New (find companies in market, not possible with Generate Insights):

curl -X POST "https://signals.autobound.ai/v1/companies/search" \
  -H "X-API-KEY: $SIGNAL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"signal_types": ["hiring-trends"], "industries": ["Software Development"], "employee_count_min": 200, "limit": 20}'

Shortened response shape (company enrich):

{
  "company": { "name": "...", "domain": "stripe.com", "linkedin_url": "...", "industries": ["..."] },
  "signals": [
    {
      "signal_id": "...",
      "signal_type": "news",
      "signal_subtype": "...",
      "signal_name": "...",
      "detected_at": "2026-10-01T00:00:00Z",
      "association": "company",
      "contact": null,
      "data": { "summary": "..." }
    }
  ],
  "signal_summary": { "news": 12 },
  "coverage": { "total_before_filters": 40, "total_after_filters": 12 }
}

Tip: contact-level coverage is lower than company-level coverage. If a contact call returns contact_not_found, call /v1/companies/enrich with the person's company domain. The 404 response lists this as a next_steps option. Entity resolution explains how we match records.

Errors you may see

The 410 response body:

{
  "error": {
    "code": 4006,
    "message": "The Generate Insights API (v1.0.0 to v1.4) was retired on 2026-10-30. Use the Autobound Signal API: POST https://signals.autobound.ai/v1/contacts/enrich or POST https://signals.autobound.ai/v1/companies/enrich. Docs: https://autobound-api.readme.io/reference/enrichcontact. The Signal API needs its own API key: https://signalapi.autobound.ai"
  }
}

Timeline

Questions

• Does my Generate Content API code change? No. v3.5 and v3.6 keep the same request, response and key.
• Do I need a new key? Yes, for the Signal API only. Get it at signalapi.autobound.ai.
• Can I use the Signal API in Claude or ChatGPT? Yes. Connect the Autobound MCP server.
• What about my plan and credits? Your account manager will help you set up Signal API credits that fit your use. See Credits, auto top-up and pay by invoice.
• Can I keep using Generate Insights after October 30? No. Every call returns 410.
• Where do I start? Introduction, then the onboarding guide.

Need help? Reply to your account manager, or email [email protected]. We are happy to do a short call to help you switch.