operations / agent

Company Brain

A Slack teammate that keeps memory of the channels it joins, answers from it and acts in connected tools, deployed to your own Cloudflare account.

At a glance

Purpose
A Slack teammate that keeps memory of the channels it joins, answers from it and acts in connected tools, deployed to your own Cloudflare account.
Runs in
Web service
Operating systems
Not stated
Needs
Runtime: Cloudflare Workers, Durable Objects and D1; Needs internet: Yes; Needs accounts or API keys beyond your agent: Yes
Files and data
Reads files: Not stated; Writes files: Not stated; Sends data to other services: Yes
Tool type
Agent
Delivery
From GitHub
Version
Latest on the main branch
Access
Code on GitHub
Compatibility
According to the README: Slack, and Cloudflare Workers with Durable Objects, D1, KV and Workers AI. Model keys can come from Anthropic, OpenAI, Google or xAI, and an OpenRouter key also works. Tools include GitHub, Linear, Notion and Google Workspace over MCP. The code sandbox can use a Daytona key, or a built-in Cloudflare container on Workers Paid. Code Mode runs in QuickJS inside the worker. Local development uses bun. The README names no operating systems.
External costs
The README requires a supermemory API key and model access through Anthropic, OpenAI, Google, xAI or OpenRouter. The README says users pay the model provider directly. The README describes Cloudflare free and Workers Paid routes, with a Daytona key for the code sandbox on the free route or a built-in container on Workers Paid. The README describes Company Brain itself as free and open source and gives no supermemory or Daytona prices. Your coding agent or model provider may charge for usage.
Permissions
The README's deploy route provisions D1, KV, Durable Objects and Workers AI in the user's Cloudflare account, using supplied supermemory and model API keys and Slack credentials. The README says the first Slack sign-in owns the deployment. The README describes shared memory for public channels, channel memory plus shared memory for private channels, and personal memory, joined private-channel memory and shared memory for DMs. The README says tool writes use the asker's connection and requests needing a teammate's connected tool receive an approve-or-deny card. The README describes MCP tool actions, sandbox repository cloning and script execution, scheduled Slack or DM digests and unprompted participation. The README's local route installs dependencies with bun and requires a tunnel for Slack access. The top-level file list includes scripts/, whose contents the README does not describe.
Usage terms
Apache License 2.0, as stated in the repository. See its LICENSE file for the exact terms.
Interface
slack

Code on GitHub

supermemoryai/company-brain

Apache-2.0

TypeScript994 stars139 forksLast pushed 2 Oct 2026 UTC

Open-sourcing our company brain - A teammate in your Slack that remembers everything your team says, and can go do the work.

Public GitHub data fetched · 0 days old

View on GitHub ↗

What it does

According to the README, Company Brain is a Slack teammate that keeps decisions, projects, owners and context from joined channels and answers from team conversations and connected tools. The README says the former paid supermemory product was discontinued and rebuilt as open-source software for the user's Cloudflare account.

The README describes actions through GitHub, Linear, Notion, Google Workspace and other MCP tools, configurable proactivity, scheduled digests and self-started research. The README says its code sandbox clones repositories, runs scripts and returns charts, CSVs and PDFs in Slack. The README describes skills and a workspace prompt for behavior, and a permissions graph for memory access.

The README identifies TypeScript, Cloudflare Workers, Durable Objects and D1, with memory on supermemory. The README describes a Slack setup page and a web home for the memory graph and settings.

Who it is for

Teams that use Slack and want to deploy their own Slack teammate that keeps team context and acts in connected tools.

Example tasks

Deploy it to your Cloudflare account and open /setup on the new worker.
Install the app to Slack and ask it a question in a channel.
Connect a tool such as GitHub, Linear or Notion over MCP.
Set how chatty it is, org-wide or per channel.
Schedule a digest to a channel or to your DMs.

Getting started

The README's main route is a deploy button for Cloudflare that asks for SUPERMEMORY_API_KEY and MODEL_API_KEY. Then open /setup on your new worker. It checks your keys, gives you a Slack app manifest and takes the Slack credentials back. Sign in with Slack, then install the app to Slack. For local development the README lists bun install, cp .dev.vars.example .dev.vars and bun run dev. The README has the exact commands.

Limitations

According to the README, deployment asks for two keys. SUPERMEMORY_API_KEY is where the brain keeps its memory. MODEL_API_KEY is an Anthropic, OpenAI, Google or xAI key, and an OpenRouter key also works. It needs a Cloudflare account and a Slack app. The README says it runs on Cloudflare's free plan, but long multi-step answers can be cut short there because the free plan allows 50 outbound calls per request. It says the code sandbox needs a Daytona key (DAYTONA_API_KEY) on the free plan, or the built-in container on Workers Paid. Local development uses bun, needs Docker if the Workers Paid block is enabled, and needs a tunnel so Slack can reach your machine. The README says the first person to sign in with Slack owns the deployment. The repository can change at any time; check its README before you install.

Data practices

According to the README, memory is kept on supermemory using your SUPERMEMORY_API_KEY, and model calls go to the provider that your MODEL_API_KEY belongs to. The README says D1, KV, Durable Objects and Workers AI are provisioned on your own Cloudflare account. The README says the brain answers from your own conversations and connected tools. Slack credentials are entered on the /setup page. The README does not mention telemetry.

Community experience

Not rated yet

Sign in to share your experience.

Comments 0

No comments yet.

Sign in to comment

Report a listing concern