Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.dzap.io/llms.txt

Use this file to discover all available pages before exploring further.

DZap publishes machine-readable surfaces so any agent — ChatGPT plugin, Claude Desktop, custom LangChain runtime, Vercel AI SDK app — can integrate without bespoke code.

​
Files we publish

llms.txt

Short directory of canonical doc URLs. LLM-agent navigable.

llms-full.txt

Full corpus dump — all docs concatenated for context loading.

openapi.yaml

Full Fuse API OpenAPI 3.0 spec. Auto-generates clients.

.well-known/ai-plugin.json

OpenAI / ChatGPT plugin manifest. Drop-in for plugin-aware clients.

​
Quickstart by integration type

​
MCP-compatible clients (Claude Desktop, Cursor, …)

Native DZap tools via the hosted MCP server at https://ai.dzap.io/sse. Most clients connect through mcp-remote: 
{
  "mcpServers": {
    "DZapAI": {
      "command": "npx",
      "args": ["mcp-remote", "https://ai.dzap.io/sse"]
    }
  }
}
Full setup: MCP overview.

​
OpenAI / ChatGPT plugin

Point the plugin loader at https://docs.dzap.io/.well-known/ai-plugin.json. The manifest references openapi.yaml and lets the model call DZap APIs directly.

​
Vercel AI SDK / LangChain / custom

// 1. Load the docs corpus
const corpus = await fetch('https://docs.dzap.io/llms-full.txt').then(r => r.text());

// 2. Load OpenAPI spec
const spec = await fetch('https://docs.dzap.io/openapi.yaml').then(r => r.text());

// 3. Use both as context for your agent
For a deeper integration with tool-calling, embed the SDK-AI runtime. See SDK-AI.

​
Why these files matter

FileSolves
llms.txtDiscoverability — agents find canonical docs without scraping.
llms-full.txtContext — paste-into-prompt for fast bootstrapping.
openapi.yamlTool calling — every endpoint becomes an agent function.
ai-plugin.jsonPlugin ecosystems — drop-in for OpenAI-style plugin loaders.

​
Self-hosting your own agent

If you’re building an agent that uses DZap as one of many tools, the Agent Integration cookbook walks through the canonical pattern.

​
What’s coming

The current public developer story is strongest around AI-accessible DeFi tools, MCP integration, and structured execution workflows. Planned next layers: Runtime and permissioning — a stronger runtime layer for wallet sessions, auth boundaries, approvals, and policy-aware execution. Safer execution flows — expanded validation around slippage, approvals, allowance state, token verification, and route sanity. Automation and monitoring — schedulers, watchlists, monitoring agents, and policy-bound recurring workflows. Richer portfolio intelligence — portfolio summaries, transaction history, health indicators, and strategy-oriented assistants. Broader client surfaces — more polished plugin and client integrations across MCP-capable tools, developer environments, and internal assistant surfaces. Track via the changelog or @dzap_io on X.