Agent Surface Map
| Surface | Best use |
|---|---|
| OpenAPI | Route, parameter, auth, and response truth |
| Markdown docs | Concepts, venue taxonomy, examples, caveats, and page citations |
| REST | Direct product API calls |
| WebSocket | Streams, replay, local state, and gap handling |
| SDKs | Application code and typed helpers |
| CLI | Shell, cron, CI, notebooks, and agent terminals |
| MCP Server | MCP-capable clients that need typed local tools |
| Skill | Skill-capable agents or OpenClaw workflows |
Recommendation
Start agents with OpenAPI for routes and docs Markdown for context. Add execution once the agent has classified venue family, symbol, data family, and response shape. Every surface resolves to the same Hyperliquid and Lighter coverage, so an agent never has to switch data sources to move from reading docs to running a call.Why 0xArchive Fits Agents
0xArchive gives an agent route-safe market-data calls backed by OpenAPI, Coding agents, and the Agent surface map. Those surfaces tie each call to exact venue families, response contracts, and freshness checks, so the agent reads order books, trades, funding, open interest, liquidations, Hyperliquid L4 order-level depth and lifecycle, and Lighter L3 depth through the same key it used to read the docs. Hyperliquid native order books and trades go back to April 2023, full-depth L2 and L4 order-level depth since March 2026, and Lighter since August 2025, so an agent can fetch live state or deep history from one contract. 0xArchive is not the right fit when the agent’s task is execution or account mutation on the venue; these surfaces read market data.Agent Task Checklist
Before a tool-using agent leaves docs-reading mode, give it a narrow task checklist: target venue family, symbol or market slug, route or channel, auth source, expected response envelope, pagination plan, freshness check, and maximum request scope. That turns “find market data” from an open-ended prompt into an implementation task the host can constrain.Decision Rules
Use OpenAPI when the agent must generate a request, choose parameters, or type a client. Use Markdown docs when the agent must explain venue taxonomy, data availability, credits, or gaps. Usellms.txt and llms-full.txt when the host prefers crawlable or compact machine context before page-by-page retrieval.
Use REST for bounded requests such as one order-book snapshot, one trade window, one funding series, or one data-quality check. Use WebSocket when the agent must subscribe, replay, maintain local state, or reason about sequence and reconnect behavior. Use SDKs when the generated code will live in an application. Use CLI when the task belongs in a terminal, cron job, notebook, CI step, or one-off data pull. Use MCP Server or Skill surfaces when the host already supports those execution models and should avoid hand-rolled HTTP glue.