Agenticfeed Standard v0.3.0 — Simplified Agent Query API

Agenticfeed Standard v0.3.0

This release simplifies the Agent Query API, removes external dependencies, and clarifies the pricing model.

What's changed

Agent Query API (breaking change)

• API now uses Agenticfeed's internal matching engine exclusively — callers no longer need to provide their own Anthropic API key
• Authentication simplified to a single Agenticfeed API key (Bearer token)
• API is now included with all paid subscription tiers (Growth, Scale, Enterprise)
• Merchants can instantly generate keys in their dashboard — no approval process

Why

• Developers integrating with Agenticfeed can now use the API without setting up external AI accounts
• Transparent pricing: no hidden per-query costs, everything included in the subscription
• Simpler onboarding and documentation
• Agenticfeed becomes a complete, self-contained platform rather than a data layer

Migration guide

If you are already using the Agent Query API with v0.2.0:

Before (v0.2.0):

{
  "query": "ergonomic chair for back pain",
  "anthropic_api_key": "sk-ant-your-key",
  "customer_guid": "a9a8378c94"
}

After (v0.3.0):

{
  "query": "ergonomic chair for back pain",
  "customer_guid": "a9a8378c94"
}

Remove the anthropic_api_key field entirely. The request now only requires your Agenticfeed API key in the Authorization header.

New Documentation

• Full API documentation with updated examples: agenticfeed.ai/api.html
• Developer blog post: Introducing the Agenticfeed Agent Query API

---

For the full standard specification, see README.md.

Agenticfeed Standard v0.2.0 — Agent Query API

Agenticfeed Standard v0.2.0

This release adds the Agent Query API — an active layer that lets AI agents query product data by natural language buyer intent, on top of the passive feed discovery introduced in v0.1.0.

What's new

Agent Query API

• POST /agent/query endpoint accepts a natural language buyer intent and returns a ranked schema.org ItemList of matched products
• Each result includes a match_reason field explaining why the product fits the buyer's need — designed to be passed directly to the buyer by the calling agent
• Matching is performed by Claude Haiku

Bring your own Anthropic key

• Callers supply their own Anthropic API key in the request body as anthropic_api_key
• The key is validated with a minimal one-token call before any product data is fetched
• Callers pay Anthropic directly at standard rates — the feed provider pays nothing for AI compute

A2A-compatible Agent Card

• Agent Card published at /.well-known/agent.json
• Describes capabilities, accepted inputs, output format, and authentication in machine-readable format
• Compatible with Google's Agent-to-Agent (A2A) protocol for automatic agent discovery

Attribution

• All product URLs in query responses are UTM-tagged (utm_source=agenticfeed&utm_medium=ai-agent)
• Merchants can measure AI agent-driven revenue in their existing analytics

Reference implementation

agenticfeed.ai — includes the query endpoint, Agent Card, and developer documentation at agenticfeed.ai/api.html

Agenticfeed Standard v0.1.0

Agenticfeed Standard v0.1.0

Initial public release of the Agenticfeed Standard — an open specification for making ecommerce product catalogues readable by AI shopping agents.

What's included

• rel="agenticfeed" discovery tag specification
• JSON-LD feed index format (DataFeed schema.org type)
• Intent endpoint formats — questions (Question), problems and use cases (ListItem)
• Product detail format (Product schema.org type with intent namespace)
• UTM attribution standard for tracking AI agent traffic
• Complete working examples for all endpoints
• llms.txt for AI crawler discovery

Reference implementation

agenticfeed.ai — the hosted implementation of this standard.