AgentWyre — AI Ecosystem Intelligence

Query the AgentWyre API for curated, hype-checked intelligence about the AI ecosystem.

When to Use

• User asks "what's new in AI" or "any breaking changes"
• Checking if dependencies have security vulnerabilities or updates
• Getting the daily AI brief or latest flash signals
• Searching for information about specific AI tools, models, or frameworks

Quick Start

# Get today's free signals (no key needed)
curl -s https://agentwyre.ai/api/feed/free | python3 -m json.tool

# Get full feed (requires API key)
curl -s -H "Authorization: Bearer $AGENTWYRE_API_KEY" https://agentwyre.ai/api/feed

# Check service status
curl -s https://agentwyre.ai/api/status

API Endpoints

Using the Helper Script

# Get latest signals (free tier)
python3 scripts/agentwyre.py signals

# Get signals in a specific language
python3 scripts/agentwyre.py signals --lang ja

# Search (requires daily/pro key)
python3 scripts/agentwyre.py search "langchain breaking change"

# Check security for specific packages
python3 scripts/agentwyre.py security openai langchain torch

# Get flash signals (pro only)
python3 scripts/agentwyre.py flash

Set AGENTWYRE_API_KEY environment variable for authenticated access. Without it, you get the free tier (full wire, 2-day delay).

MCP Server

For deeper integration, install the MCP server:

{
  "mcpServers": {
    "agentwyre": {
      "command": "npx",
      "args": ["agentwyre-mcp"],
      "env": { "AGENTWYRE_API_KEY": "your_key" }
    }
  }
}

Tiers

• Free — Full daily wire, 2-day delay. No key needed.
• Daily ($2.99/mo) — Same-day access, search, 8 languages.
• Pro ($9.99/mo) — Everything + hourly flash, security advisories, cost tracking.
• USDC on Base accepted — no international bank fees.

Interpreting Signals

Each signal includes:

• title + summary — What happened
• category — release, security, ecosystem, infrastructure, research, breaking_change
• severity — critical, high, medium, low
• confidence — 1-10 (single source capped at 6)
• hype_check — hype_level (verified/promising/overhyped/vaporware), reality, red_flags, green_flags
• action — Recommended steps with commands and rollback instructions
• sources — Verifiable URLs

Always present actions to the user before executing. All actions include requires_user_approval: true.

More info: https://agentwyre.ai/faq