Feedback for Claude, ChatGPT & Cursor

by B77 Productivity

0.0 0 reviews installs Updated 18d ago

Every feature request and bug from your chat channels — captured, categorized, countable.

Turn #feature-requests and #bugs into a structured product backlog.

Works with ClaudeChatGPTCursor

About

What it does.

Connect your team's feature-request and bug channels in Slack or Mattermost. Every message is saved word-for-word, structured by AI into a titled, categorized record, matched against similar past requests, and confirmed back in the channel. When you plan, ask what users wanted most — with real counts.

A memory for everything your users ask for:

- Connect a feature-requests channel and a bugs channel from Slack or Mattermost in one step each — you get a webhook address to paste, and confirmations appear right in the channel. - Every post is preserved verbatim and structured by AI: a clear title, summary, category, and severity for bugs or demand signal for feature requests. - Similar requests are detected automatically, so you see "asked 7 times" instead of seven scattered messages. - Product decisions live on the record: set a status (planned, rejected, fixed…) and leave a decision comment your team can always find. - When planning, search the backlog by meaning, not keywords — "what did salons ask about calendars?" just works. - Link any item to a Teamwork task so the request that motivated a ticket is one hop away.

Built and operated by B77.

What you can ask

Just type and go.

What did users ask for most this month?

Show new critical bugs

Mark this request as planned and note why

Find every request about calendar sync

Capabilities

Read + write.

Your assistant can look things up and, when you ask, create or change data in the connected tool.

Works in Claude, ChatGPT, Cursor · hosted at feedback.b77.ai/mcp · what this app can see · how access is secured

Reviews

What people say.

0.0

0 reviews

Write a review

0.0

0 reviews

5
0%
4
0%
3
0%
2
0%
1
0%

What's new

Version 1.0.0

18 days ago

Version history
  • Initial release — Slack/Mattermost channel capture, AI categorization with severity and demand signals, similarity counting, statuses with decision comments, semantic search.

Q & A

Questions, answered.

How do messages get in?

You connect each channel once: the app gives you a webhook address, you paste it into your Slack or Mattermost channel settings, and every new post in that channel is captured automatically.

Does it change what people wrote?

Never. The original message is stored word-for-word forever. The AI adds a structured layer on top — title, summary, category, severity — for planning and search.

How does it help prioritize?

Similar requests are counted together, bugs carry a severity, and feature requests carry a demand signal (sales blocker, churn risk, explicit request). Ask for the top-requested items over any period and plan from evidence.

Can I act on an item from the chat?

Yes. Set statuses, leave a product decision comment, or link the item to a Teamwork task — all by asking your AI.

How do I connect Feedback MCP to Claude?

Feedback is a hosted MCP server, so it becomes a Claude connector without installing anything locally. 1. Open the app page on B77 and click Install — copy the connection URL it shows. 2. In Claude: main menu → Customize → Connectors → Add → Add custom connector. 3. Name it after what it does, paste the URL into "Remote MCP server URL", sign in with your B77 account. The same connector shows up in claude.ai, Claude Desktop and Claude Code. The connection URL is https://feedback.b77.ai/mcp.

How do I connect Feedback MCP to ChatGPT?

Feedback is a hosted MCP server, so it becomes a ChatGPT plugin without installing anything locally. 1. Open the app page on B77 and click Install — copy the connection URL it shows. 2. In ChatGPT: Settings → Safety & Login → enable Developer Mode. 3. Open Plugins → "+" top right → name the plugin, paste the B77 URL, sign in with your B77 account — it is now available in any chat. The connection URL is https://feedback.b77.ai/mcp.

How do I connect Feedback MCP to Cursor?

Feedback is a hosted MCP server, so it becomes a Cursor MCP server without installing anything locally. 1. Open the app page on B77 and click Install — copy the connection URL it shows. 2. In Cursor: Settings → MCP servers → "+ Add new MCP server" — or edit mcp.json directly: { "mcpServers": { "teamwork": { "url": "https://teamwork.b77.ai/mcp" } } }. 3. Save and restart; Cursor asks you to sign in with your B77 account once, then the tools are available to the agent. The connection URL is https://feedback.b77.ai/mcp.

What data can Feedback access?

Reads: Messages posted in the specific channels you connect — nothing else from your chat workspace. Stores: The original message text, its author's username, and the AI-structured record built from it. Uses: AI categorization runs under a fixed daily budget; the original text is always kept unchanged. Access runs through your own OAuth sign-in and stays scoped to your account — B77 never sees your passwords, and you can disconnect the app at any time.

Is Feedback free?

Yes. Feedback is free on B77 — no card, no trial clock. Install it, connect it to your assistant and use it. If Feedback talks to a third-party service, that service's own plan still applies.

Can I share Feedback with my team?

Yes. From the app's Manage page on B77 you invite teammates by email; they add the same Feedback connection to their own assistant and use it under your access, without setting up credentials themselves. Team apps such as Knowledge Base, Teamwork and Feedback also let you invite people per project. You can revoke a share at any time, and every call is written to the usage log.

Privacy

What this app can see.

Privacy

GDPR-aligned.

Right-to-be-forgotten honoured. No silent retention.

Residency

EU-hosted.

Frankfurt + Helsinki regions. Data stays in the bloc.

Auditability

Audit-log on.

Every tool call logged. SOC2-ready exports.

Reads
Messages posted in the specific channels you connect — nothing else from your chat workspace.
Stores
The original message text, its author's username, and the AI-structured record built from it.
Uses
AI categorization runs under a fixed daily budget; the original text is always kept unchanged.