Decixa Agent Hub

Two ways to use Decixa

Both flows end the same way: pay per call with x402. No accounts, no API keys.

Both flows sit on the same Trust Layer — every listed API is verified daily (uptime, latency, schema). The only difference is who does the comparison: your agent (discover) or Decixa (resolve).

Quick Start

# 1. Fetch the manifest to discover endpoints
curl https://api.decixa.ai/.well-known/agent-index.json

# 2a. POST an intent and get the top recommendation
curl -X POST https://api.decixa.ai/api/agent/resolve \
  -H "Content-Type: application/json" \
  -d '{"intent":"web scraping","constraints":{"latency_p95_max_ms":500,"cost_max_per_call_usdc":0.05}}'

# 2b. Or list multiple candidates with filters
curl "https://api.decixa.ai/api/agent/discover?task=web+scraping&latency_p95_max_ms=500&cost_max_per_call_usdc=0.05"

# 3. Get full details for a specific API
curl "https://api.decixa.ai/api/agent/detail/{id}"

Use via MCP Server

The fastest way to use Decixa from Claude Code or any MCP-compatible client.

# 1. Start the MCP server
$ npx decixa-mcp

# 2a. Get the top recommendation for an intent
mcp: resolve({ intent: "extract social media posts by keyword" })

# 2b. Or list multiple ranked candidates so the agent can choose
mcp: discover({ intent: "extract social media posts", limit: 5 })

# → Decixa returns ranked results with cost, latency, and trust signals

Use via TypeScript SDK

Build Decixa into your agent with a few lines of code.

$ npm install decixa-sdk

import { Decixa } from "decixa-sdk";
const decixa = new Decixa();

// resolve — single recommendation
const { recommended } = await decixa.resolve({
  intent: "extract social media posts by keyword",
});

// discover — ranked list (paginated)
const { apis } = await decixa.discover({
  intent: "extract social media posts",
  limit: 5,
});

Use via CLI

Search APIs directly from your terminal.

# resolve — single recommendation
$ npx decixa resolve "extract social media posts by keyword"

# search — ranked list (alias for discover)
$ npx decixa search "extract social media posts"

  Found 10 APIs for "extract social media posts"

  1. StableSocial — posts (Capability: Extract  |  Price: $0.06 USDC)
  2. ...

Endpoints

curl https://api.decixa.ai/.well-known/agent-index.json

{
  "name": "Decixa",
  "tagline": "The Index for Agents",
  "discover_endpoint": "https://api.decixa.ai/api/agent/discover",
  "resolve_endpoint": "https://api.decixa.ai/api/agent/resolve",
  "pagination": { "default_limit": 20, "max_limit": 50 },
  "pricing": { "access": "free" }
}

curl -X POST https://api.decixa.ai/api/agent/resolve \
  -H "Content-Type: application/json" \
  -d '{
    "intent": "scrape reddit posts",
    "constraints": {
      "latency_p95_max_ms": 500,
      "cost_max_per_call_usdc": 0.01
    }
  }'

{
  "recommendation_status": "resolved",
  "recommended": {
    "id": "...",
    "name": "Reddit Sentiment API",
    "endpoint": "https://api.example.com/reddit",
    "pricing": { "model": "per_call", "usdc_per_call": 0.002 },
    "ranking_basis": "vector_similarity + latency + price + tiebreak"
  },
  "alternatives": [ /* up to 2 */ ]
}

curl "https://api.decixa.ai/api/agent/discover?task=scrape+reddit&latency_p95_max_ms=500&cost_max_per_call_usdc=0.01"

{
  "total": 42,
  "limit": 20,
  "offset": 0,
  "next": "https://api.decixa.ai/api/agent/discover?offset=20",
  "apis": [
    {
      "id": "...",
      "name": "Reddit Sentiment API",
      "pricing": { "model": "per_call", "usdc_per_call": 0.002 },
      "agent_ready": true,
      "latency_tier": "low",
      "detail_url": "https://api.decixa.ai/api/agent/detail/..."
    }
  ]
}

curl "https://api.decixa.ai/api/agent/detail/{id}"

{
  "id": "...",
  "name": "Web Crawl API",
  "endpoint": "https://api.example.com/crawl/v1",
  "agent_compatibility": {
    "agent_ready": true,
    "latency_tier": "high",
    "execution_mode": "async",
    "deterministic": false
  },
  "schema": {
    "input_type": ["url"],
    "output_type": ["json", "html"],
    "spec_url": "https://api.example.com/openapi.json"
  },
  "use_cases": ["web scraping", "content extraction"]
}

Filters & Constraints

Roadmap