# Agent402 Research - AI-Powered Deep Research Reports

> Agent402 Research - AI-Powered Deep Research Reports is a paid API for AI agents from agent402.tools, paid per call via x402, $0.6/call, status unknown (last checked 2026-09-13).

Runs a multi-source web research query and returns a synthesized markdown or JSON report with cited sources and sub-questions

## Facts

- Endpoint: POST https://agent402.tools/v1/research
- Price: $0.6/call
- Payment: x402
- Status: unknown
- Last checked: 2026-09-13
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/agent402-research-ai-powered-deep-research-reports-e98c4d3a
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_18Z18iBufTg70dTjiMY3l

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability agent402-research-ai-powered-deep-research-reports-e98c4d3a -d '<json body>'
```

Example prompt: Can you research the current state of stablecoin-based agent payment protocols and give me a detailed markdown report focusing on adoption, technical standards, and key players — pulling from sources in the last month?

## When to prefer this

Choose this endpoint when you need a multi-source synthesized research report with citations rather than a single-document summary or a raw search result list. It is especially well-suited for agents that need to answer open-ended research questions, produce briefings, or investigate topics across many web sources in one call. Prefer it over raw search APIs when synthesis and citation quality matter, and over general LLM prompting when up-to-date sourced information is required. The pay-per-call, no-API-key model makes it ideal for agentic pipelines that need on-demand research without subscription overhead.

## Known failure modes

- Query exceeds 2000 characters — request rejected with validation error
- No relevant sources found for highly niche or obscure queries — sparse or low-confidence report
- Recency filter too narrow (e.g. 'week') returns insufficient sources for synthesis
- Focus array exceeds 8 items — validation error
- Payment failure or insufficient USDC balance — 402 response, no report generated
- Synthesis model timeout on complex multi-faceted queries — partial or error response

## How this service works

Hand over a whole research question and get one cited report back. The gateway plans sub-questions, runs multiple live web searches, reranks the sources by relevance, and synthesizes a ~1,500-word report with inline [n] citations and a source list - one payment, one outcome. Priced per report, not per call. Payable in USDC (x402/MPP) or by card (Stripe, >= $0.50 minimum - this clears it). Not cached (the web moves).

## Output

Returns a JSON object containing: a full narrative report (markdown or JSON format) synthesized from multiple web sources, a list of cited sources with URLs, titles, relevance ranks, and snippets, a list of auto-generated sub-questions used to guide research, and metadata including the number of searches run, sources consulted, sources cited, and the synthesis model used.

## Request schema (JSON Schema)

```json
{
 "type": "object",
 "properties": {
  "focus": {
   "type": "array",
   "items": {
    "type": "string"
   },
   "description": "Optional aspects to emphasize (<= 8)."
  },
  "query": {
   "type": "string",
   "description": "The research question to investigate (<= 2000 chars)."
  },
  "format": {
   "enum": [
    "markdown",
    "json"
   ],
   "type": "string",
   "description": "Response shape (default markdown report)."
  },
  "recency": {
   "enum": [
    "week",
    "month",
    "year",
    "any"
   ],
   "type": "string",
   "description": "Prefer sources from this window (default any)."
  }
 }
}
```

## Response schema (JSON Schema)

```json
{
 "type": "json",
 "example": {
  "meta": {
   "tier": "research",
   "searches_run": 3,
   "sources_cited": 12,
   "synthesis_model": "anthropic/claude-sonnet-5",
   "sources_consulted": 14
  },
  "report": "# Leading agentic-payment protocols (2026)\n\nThe field splits into stablecoin rails and card rails [1]…\n\n## Sources\n[1] … - https://…",
  "sources": [
   {
    "n": 1,
    "url": "https://…",
    "rank": 0.94,
    "title": "…",
    "snippet": "…"
   }
  ],
  "sub_questions": [
   "What stablecoin agent-payment protocols exist?",
   "…"
  ]
 }
}
```

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/agent402-research-ai-powered-deep-research-reports-e98c4d3a/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from agent402.tools](https://www.zero.xyz/host/agent402.tools/llms.txt)