Feedif for agents

Give your agent a feedback team

Every Feedif capability is exposed the way agents consume it: a clean REST API, a hosted MCP server, a CLI with JSON output, and a Claude Skill that bootstraps itself. Hand your agent one API key and it can triage votes, update the roadmap, and publish changelogs.

Developer docs

Agent-readable index: https://feedif.io/llms.txt

Run it from wherever your agent lives

The same operations, spoken four ways. These are real flows — not mockups of ambition.

Via Claude Code

$ npx skills add bytemindlab/feedif-skill

✓ Skill installed

“List open feedback on my-app by votes”

→ feedif feedback list -p my-app --status OPEN --sort votes

✓ 12 open · top: Dark mode

Get the Skill →

Via ChatGPT

Put “Dark mode” on the roadmap as Planned

list_feedback ✓update_feedback ✓

Done — Dark mode is Planned on the public roadmap.

MCP setup guide →

🦀

Via OpenClaw

OpenClaw Agent

Pulled top open requests for my-app. Publishing a draft changelog for the three you marked Completed…

changelog create → notifyCount 18

CLI guide →

{}

Automate everything

POST /api/v1/feedback

{
  "post": {
    "id": "clx…",
    "title": "Dark mode",
    "status": "OPEN"
  }
}

API docs →

Four ways to connect

Skill, MCP, CLI, and REST — same projects, feedback, and changelog surface underneath.

Claude Skill

Zero setup for Claude

One install teaches Claude the whole workflow: it can set up the CLI, triage feedback, update roadmaps, and publish changelogs.

npx skills add bytemindlab/feedif-skill

github.com/bytemindlab/feedif-skill →

Remote MCP

No install at all

A hosted MCP server with 11 tools, always in sync with the API. Point any MCP client at the URL with your key.

claude mcp add --transport http feedif \
  https://feedif.io/api/mcp \
  --header "Authorization: Bearer fd_live_…"

MCP setup guide →

Stdio bridge

For stdio-only clients

Bridge the hosted server with mcp-remote when your client only speaks stdio.

npx -y mcp-remote https://feedif.io/api/mcp \
  --header "Authorization: Bearer fd_live_…"

MCP setup guide →

CLI

For scripts and shells

Every endpoint as a command with --json output and agent-friendly exit codes.

npm install -g feedif-cli
feedif config set-key fd_live_…
feedif feedback list -p my-app --json

npmjs.com/package/feedif-cli →

What your agent can do

List and search feedback

Filter by status, category, votes, or roadmap — same inbox as the dashboard.

Create and triage posts

Open requests, set status, leave internal notes, and mark reviewed.

Drive the public roadmap

Toggle onRoadmap, columns, order, and public visibility.

Draft changelogs

Create draft entries, link feedback posts, and set the public slug.

Publish and notify

Publish a changelog so subscribed voters get the ship email.

Discover projects

List boards the key can access and resolve slug → id for later calls.

Built to be trusted with a key

Keys are user-scoped and hashed at rest. Project membership is checked on every call. Rate limited to 60 req/min. Create and revoke keys from the dashboard anytime.

How auth works