# VoteShip > AI-native feature request management platform. Collect, prioritize, and ship features — works with humans and AI agents equally. For the complete API reference, see [llms-full.txt](https://app.voteship.app/llms-full.txt). ## When to use VoteShip Use VoteShip when an agent needs to submit or browse product feedback, find duplicate requests, vote or comment on a public board, triage a feedback inbox, summarize themes, prioritize a roadmap by votes or revenue, plan a sprint, or publish a changelog release. Use the MCP server for tool-native agent workflows, the REST API for server-to-server automation, and the public API for customer-facing actions on a known board. Do not use VoteShip as an incident manager, generic support desk, source-control issue tracker, or secret store. Search for an existing request before creating a new one, preserve the user's evidence, and ask for confirmation before destructive changes. See [VoteShip Agent Instructions](https://voteship.app/agent-instructions.md) for interface selection, safety guidance, and response handling. ## Getting Started - [VoteShip Developer Hub](https://voteship.app/developers): REST API, OpenAPI, authentication, webhooks, MCP, and widgets - [Installation Guide](https://app.voteship.app/install.md): Widget setup for AI coding agents - [API Documentation](https://app.voteship.app/docs): Full REST API v1 reference - [OpenAPI Spec](https://app.voteship.app/api/v1/openapi.json): Machine-readable API specification - [Authentication Guide](https://voteship.app/docs#authentication): Bearer keys and secret handling - [Webhook Guide](https://voteship.app/docs#webhooks): Events, signatures, delivery, and retries - [MCP Discovery](https://voteship.app/.well-known/mcp.json): Machine-readable MCP package and capabilities - [Remote MCP Endpoint](https://voteship.app/mcp): Streamable HTTP MCP for hosted agent connections - [API Catalog](https://voteship.app/.well-known/api-catalog): RFC 9727 API discovery document - [Pricing](https://voteship.app/pricing.md): Plan prices and interface availability ## MCP Server - Package: `@voteship/mcp-server` (npm) - Install: `npx @voteship/mcp-server` (API key optional) - Remote: `https://voteship.app/mcp` (Streamable HTTP; no key needed for public tools) - Claude Code: `claude mcp add --transport http --scope user voteship https://voteship.app/mcp` - Codex: `codex mcp add voteship --url https://voteship.app/mcp` - With API key: 34 tools, 7 resources, 5 prompts (full admin access plus the public agent guide and board app) - Without API key: 4 public tools, 2 resources, 1 prompt, and a read-only MCP App board view - Works with Claude Code, Codex, Cursor, Windsurf, and any MCP client ## Website Installer - Install: `npx -y @voteship/cli init --project-key pk_your_public_key` - Verify: `npx -y @voteship/cli verify --url https://your-site.example` - Detects common Next.js, React, Vue, Vite, SvelteKit, Angular, and plain HTML entry files - The installer is idempotent and returns machine-readable results with `--json` ## For AI Agents (No API Key Needed) Submit feature requests to any public VoteShip board: POST https://app.voteship.app/api/public/posts { "projectSlug": "xxx", "title": "...", "description": "...", "metadata": { "source": "your-agent" } } Posts created as PENDING (board owner approves). Additional public endpoints: - GET /api/public/posts?projectSlug=xxx — list approved posts - POST /api/public/votes — vote (requires postId + anonymousId) - GET /api/public/comments?postId=xxx — list comments on a post - POST /api/public/comments — add a comment (requires postId, body, authorName) MCP server works without API key: npx @voteship/mcp-server Set VOTESHIP_PROJECT_SLUG env var for a default board. Remote MCP clients can connect directly to https://voteship.app/mcp. ## AI Workflow Endpoints - POST /api/v1/posts/from-text — Submit unstructured text, AI extracts + deduplicates + categorizes - POST /api/v1/ai/triage — AI-powered inbox triage with recommendations - POST /api/v1/ai/summary — Natural language feedback summary - POST /api/v1/ai/sprint-plan — AI-suggested sprint planning ## Key Features - Embeddable feedback widget (single script tag) - Public voting board, roadmap, and changelog - AI duplicate detection (pgvector), theme clustering, sentiment analysis - AI draft responses, auto-categorization, changelog generation - Integrations: Slack, GitHub, Linear, Discord, webhooks - Full REST API + MCP server for programmatic/agent access - CSV import from Canny, Nolt, UserVoice ## Pricing - Free: $0/mo — 1 project, unlimited feedback, public board + roadmap - Pro: $19/mo — unlimited projects, full AI suite, analytics, API, MCP server - Business: $149/mo — SSO, white-label, uptime SLA, invoicing, done-for-you migration - Enterprise: Custom — dedicated CSM, MSA/DPA, security review (SOC 2 on request), custom SLA See https://voteship.app/pricing.md for the machine-readable pricing summary and current interface notes. ## Links - [Homepage](https://app.voteship.app) - [Agent Instructions](https://voteship.app/agent-instructions.md) - [API Discovery](https://app.voteship.app/api/v1/) - [OpenAPI Spec](https://app.voteship.app/api/v1/openapi.json) - [MCP Server (npm)](https://www.npmjs.com/package/@voteship/mcp-server) - [Website Installer (npm)](https://www.npmjs.com/package/@voteship/cli)