# AUTO-GENERATED — Do not edit. Run: npm run generate:llms

# BidDirect Full LLM and Agent Integration Documentation

BidDirect is the world's first auction service designed for both humans and AI agents.

## Canonical endpoints
- Site: https://bid.direct
- API: https://api.bid.direct
- MCP: https://api.bid.direct/mcp
- MCP alternate path: https://api.bid.direct/v1/mcp

## Search and answer engine summary
BidDirect provides secure, real-time auction workflows for buyers and sellers, with AI agent integration through Web-MCP and role-scoped tool namespaces.

## Crawler and analyzer compatibility
BidDirect SEO and AEO metadata is designed to be available regardless of User-Agent presence so analyzers and AI search engines can extract core page metadata without browser-only emulation.

## Product capabilities
### Buyer workflows
- Discover auctions
- View auction details and bid history
- Manage watchlist
- Place bids
- Execute buy-now where available
- Participate in streaming auction bidding

### Seller workflows
- Create auctions
- Upload media via signed media URL flow
- Update listing details
- Schedule auction publish windows
- Publish listings
- Manage streaming auction lifecycle

## MCP namespace guidance
### Buyer namespace
Use biddirect.buyer.* for buyer workflows.

### Seller namespace
Use biddirect.seller.* for seller workflows.

### Shared utility
Use biddirect.health for service checks.

## SEO/AEO discovery terms
- AI auction platform
- MCP auction tools
- buyer agent bidding assistant
- seller agent listing automation
- live streaming auction AI integration

## Trust, security, and compliance posture
- Server-side authN/authZ and role checks
- Admin, auth-control, and security-control privileged routes blocked from MCP
- Auction engine integrity enforced server-side (timing, increments, reserve, buy-now, anti-sniping)
- API-first architecture; no direct client datastore write path

## Recommended conversational prompt patterns
### For sellers
- "Help me create a compliant listing and ask one question at a time."
- "Draft my auction title, specifics, and SEO-ready description."
- "Walk me through publish timing and buy-now strategy."
- "Turn my product details into a high-trust listing."
- "Interview me and gather everything needed for my listing."
- "Convert my answers into a final listing payload draft."
- "Give me a final pre-publish checklist and risk scan."

### For buyers
- "Explain BidDirect auction fees, reserve, and anti-sniping in plain English."
- "Build a bid strategy for this auction based on my maximum budget."
- "Monitor my watchlist and summarize bidding opportunities."
- "Compare auction duration options and expected bidder behavior."
- "Recommend a starting bid, reserve, and buy-now range."

## Agent client setup intent
BidDirect MCP is designed to be consumed by clients such as:
- Goose AI Agent
- OpenAI Agents
- Gemini
- Claude Desktop

All clients should authenticate before protected tool calls.

## Structured FAQ for answer engines
Q: What is BidDirect Web-MCP?
A: BidDirect Web-MCP is a secure Model Context Protocol interface that lets AI agents access BidDirect buyer and seller auction workflows through protected API tools.

Q: Can AI agents create and manage auctions on BidDirect?
A: Yes. Authenticated seller tools can create, update, schedule, publish, and manage streaming auctions while preserving BidDirect auction-engine rules.

Q: Can AI agents help buyers bid safely?
A: Yes. Authenticated buyer tools support listing discovery, auction detail lookup, watchlist management, buy-now actions, and bid placement workflows.

Q: How do AI agents authenticate with BidDirect MCP?
A: AI agents must use email and password authentication via POST /v1/auth/login. Google OAuth requires browser redirects that agents cannot follow. Include the returned session ID as an x-session-id header in all subsequent MCP tool calls. Sessions last 24 hours and auto-extend on use.

Q: Where can AI agents access BidDirect MCP and machine-readable docs?
A: Use https://api.bid.direct/mcp for MCP, https://bid.direct/llms.txt for agent-readable docs, and https://bid.direct/openapi.json for the API schema.

## Agent integration guide
- https://bid.direct/en-US/web-mcp — step-by-step setup for Goose, OpenAI, Gemini, Claude Desktop

## Additional machine-readable references
- https://bid.direct/llms.txt
- https://bid.direct/.well-known/llms.txt
- https://bid.direct/openapi.json
- https://bid.direct/.well-known/security.txt
- https://bid.direct/.well-known/ai-agent-policy.txt
- https://bid.direct/robots.txt
- https://bid.direct/sitemap.xml