> ## 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.

# FAQ

> Common questions about DZap, Fuse, and the AI stack.

<AccordionGroup>
  <Accordion title="What is DZap?">
    DZap is a DeFi aggregation protocol. Instead of integrating DEXs and bridges one by one, you integrate DZap and route across all of them. Three surfaces: **Core** (swap + bridge aggregator), **Fuse** (single-click intent execution), and the **AI stack** (CLI, MCP, Skills, Tools).
  </Accordion>

  <Accordion title="Do I need an API key?">
    Not to start. The Fuse API (`zap.dzap.io`) is public and rate-limited per-IP. The Trade API (`api.dzap.io`) accepts optional API keys for higher rate limits and per-route configuration. See [Authentication](/api/authentication).
  </Accordion>

  <Accordion title="What's the difference between DZap Core and DZap Fuse?">
    **Core** is route-and-execute — you ask for a quote, sign a transaction, send it. **Fuse** is intent-based — you sign an intent describing the outcome you want, and a solver settles it (gasless variants supported via EIP-712 permit). See [Product Stack](/products/product-stack).
  </Accordion>

  <Accordion title="Is the AI execution safe? What signs my transactions?">
    The AI stack never holds your private key off-session. Tools that call execution paths (e.g. `PerformZapTool`) use a session-scoped key that lives only in process memory. Confirmations and limits are configurable. See [ZapBot Safety](/ai/zapbot/safety).
  </Accordion>

  <Accordion title="Why is the code called 'Zap' but the docs call it 'Fuse'?">
    Fuse is the public-facing brand for the intent execution layer. Code, repos, and SDK methods still use `Zap*` names — the rename is gradual. When you see `getZapQuote`, `signGaslessDzapUserIntent`, etc. in the SDK, that's Fuse.
  </Accordion>

  <Accordion title="Which chains are supported?">
    Most major EVM chains plus Solana and Bitcoin (for bridge endpoints). The full live list comes from `getAllSupportedChains()` in the SDK and the `/v1/config/chains` endpoint. See [Supported Chains](/products/supported-chains).
  </Accordion>

  <Accordion title="How do I report a bug or request a feature?">
    Open an issue on the relevant repo or ping us in [Discord](https://discord.com/invite/aRHESJB9md). For security issues, email `support@dzap.io` directly — see [Security](/resources/security).
  </Accordion>
</AccordionGroup>

## DZap AI

<AccordionGroup>
  <Accordion title="What is DZap AI?">
    DZap AI is the developer-facing AI layer for connecting LLMs, agents, and applications to DZap capabilities — market data, DeFi context, routing, and execution workflows. See [AI Overview](/ai/overview).
  </Accordion>

  <Accordion title="Is DZap AI only for chatbots?">
    No. Chat is one interface. The goal is to support structured tool use across assistants, apps, MCP clients, internal copilots, and workflow automation.
  </Accordion>

  <Accordion title="Why use MCP?">
    MCP gives developers a standard way to expose structured tools to compatible AI clients. It reduces custom glue code and makes tool discovery cleaner. Connect via `https://ai.dzap.io/sse`. See [MCP Server](/ai/mcp/overview).
  </Accordion>

  <Accordion title="Can DZap AI execute transactions?">
    It can support execution-aware workflows, but sensitive actions are always gated by wallet context, confirmation, and runtime policy. See [Safety & Execution](/ai/safety-and-execution).
  </Accordion>

  <Accordion title="Should I start with execution?">
    Usually not. Start with read-only and simulation flows first, then add policy-aware execution once your tool contracts and safety checks are stable. See [Getting Started](/ai/quickstart).
  </Accordion>

  <Accordion title="What tools are most useful first?">
    Most teams start with: token lookup, price and balance tools, route or quote generation, transaction building, simulation. See [Tool Categories](/ai/tools/overview).
  </Accordion>

  <Accordion title="Is DZap AI useful for internal teams?">
    Yes. A curated DZap AI setup is useful for support, ops, research, and developer enablement — not just end-user chat experiences.
  </Accordion>
</AccordionGroup>