Your feedback board, in your AI tools
Connect Claude, Cursor, Codex, or Gemini to fdback over the Model Context Protocol. Triage feedback, draft changelogs, and summarize what users want — from wherever you already work.
> draft a changelog for what we shipped this week
⏺ list_feedback — status: completed, this week
4 shipped posts across 2 boards, 137 voters
⏺ create_changelog — draft
Draft ready: “This week in fdback” — bulk actions, board search, two fixes. Publish and notify voters?
> yes, publish it
⏺ publish_changelog ✓ — 137 voters notified
One sentence runs the whole loop
Real workflows our users hand to their agents.
“Close anything that looks like spam, tag the rest by area, and move real requests to In Review.”
A weekend of submissions triaged before standup — without opening the dashboard.
“Anything completed this week without a changelog yet? Draft one in our voice and link the shipped posts.”
A draft in seconds; on publish, every voter gets notified automatically.
“What are users asking for most this quarter? Summarize the top themes with vote counts.”
A demand summary with real numbers, ready to paste into your planning doc.
Connected in one command
It's a remote server — nothing to install or host. Create an API key under Settings → API, run one command, and the workspace tools appear in your client.
Setup guides for every clientClaude Code
claude mcp add --transport http fdback \
https://app.fdback.io/api/mcp \
--header "Authorization: Bearer YOUR_API_KEY"Pick the agent you already live in
Claude CodeOne terminal command — tools appear in your next session.Claude DesktopConnects through the mcp-remote proxy in claude_desktop_config.json.
CursorDrop the server into ~/.cursor/mcp.json or per-project.
CodexAdd the server to ~/.codex/config.toml.
GeminiAdd the httpUrl entry to ~/.gemini/settings.json.What your AI can do with it
Triage in plain language
The model reads your real boards, finds duplicates, updates statuses and tags — one sentence instead of twenty clicks.
Draft changelogs from shipped work
Point it at what you completed and it writes the entry in markdown. Publish when you're happy — voters get notified.
Reply and follow up
Read threads, post comments, vote, and mark requests shipped without leaving your editor or chat.
Safe by design
No delete, archive, or merge tools exist on the server. The API key acts as you — your workspace role bounds every call.
Questions? We've got answers
Everything you need to know about collecting feedback and building your roadmap.
Contact usWhat makes fdback different from Canny, Nolt?
fdback.io is built for teams who ship fast and don't want to babysit another tool. Setup takes minutes, pricing is transparent, and you won't outgrow it on day one. No demo calls, no sales process — just sign up and start collecting feedback.
Why is fdback so affordable?
No VC funding, no enterprise sales team, no bloated overhead. Just one developer focused on building a great product at a fair price.
Can I use my own domain?
Yes. Connect your custom domain and your feedback portal lives at feedback.yourdomain.com.
How do I collect feedback from my users?
Share your feedback board link or embed the widget on your site. Users can submit ideas, vote, and comment.
How do I notify users when I ship their feature?
Mark feedback as completed and everyone who voted gets an email automatically. You can also link it to a changelog entry for a public announcement.
What is the Model Context Protocol (MCP)?
MCP is an open standard that lets AI assistants securely connect to external tools and data. fdback's MCP server exposes your feedback workspace so AI tools like Claude, Cursor, Codex, and Gemini can read and manage it on your behalf.
Do I need to install or host anything?
No. It's a remote Streamable HTTP server at app.fdback.io/api/mcp — you add the URL and your API key to your client's config and you're done. No SDK, no local process, no bridge.
Which AI tools work with fdback's MCP server?
Any MCP client that supports remote HTTP servers: Claude Code, Claude Desktop (via mcp-remote), Cursor, Codex, Gemini, and more. Setup guides for each are in the docs.
Can the AI delete things in my workspace?
No. The server deliberately exposes no delete, archive, or merge tools. Writes are limited to creating and editing feedback, comments, votes, and changelog entries — and every call runs under your workspace role.
MCP or the REST API — which should I use?
MCP is for AI assistants working conversationally; the REST API is for your own integrations and scripts. They share the same API keys and permissions, so you can use both.
What plan do I need for MCP access?
MCP access is part of the Pro plan ($15/month), which includes API access. Create a key under Settings → API and you're connected in a minute.
Hand your board to your agent
Grab an API key, paste the endpoint into your client, and ask your agent to do something it couldn't do an hour ago.