CodeRabbit Slack Agent
- CodeRabbit launched 'CodeRabbit Agent' for Slack to capture tribal knowledge from Slack, pull requests, and issue trackers. - The company claims the agent reviews over two million pull requests weekly and surfaces searchable context for teams. - It aims to reduce context loss in distributed engineering workflows by making historical decisions and rationale discoverable (x.com).
CodeRabbit has launched a Slack agent that turns engineering chat into a shared memory for code, tickets, and past decisions. (docs.coderabbit.ai) The product entered early access on April 17, 2026, and CodeRabbit said on April 22 that it is expanding from pull-request review into a broader Slack-based workflow. The company said the agent can investigate issues, generate implementation plans, and open pull requests from a Slack thread. (docs.coderabbit.ai) (adtmag.com) In CodeRabbit’s documentation, the Slack agent responds to `@coderabbit` mentions and slash commands in channels, threads, and direct messages. It can connect Slack conversations to Jira, Linear, Notion, Sentry, Datadog, PagerDuty, Figma, Google Drive, and Model Context Protocol servers. (docs.coderabbit.ai) The pitch is that software teams lose context when decisions live in chat, bug trackers, and pull requests instead of one searchable record. CodeRabbit says its agent keeps a persistent knowledge base so later conversations can reuse what a team already learned. (coderabbit.ai) (docs.coderabbit.ai) That problem has become more visible as companies push developers to use coding agents that work in short-lived sessions. CodeRabbit argued this week that engineers often spend time re-explaining architecture, old tradeoffs, and deprecations before an agent can make useful changes. (coderabbit.ai) CodeRabbit says the system is built on the same context engine that already runs about 2 million code reviews a week across 15,000 engineering teams. The company’s docs describe that context as a mix of repository rules, linked repos, issues, past pull requests, external docs, and team “learnings.” (tmcnet.com) (docs.coderabbit.ai) The Slack version adds controls meant to limit what the agent can see. CodeRabbit says each response runs inside a channel-specific “Scope,” so admins can choose which repositories, tools, and settings are available in a given Slack channel or direct message. (docs.coderabbit.ai) The company is pushing the product as a team tool, not just a personal assistant. Its docs also describe shared sandboxes for preparing code changes, recurring Slack automations for audits and summaries, and web-based thread reviews with saved context and citations. (docs.coderabbit.ai) CodeRabbit is a venture-backed startup founded in 2023 by Guritfaq Singh, Harjot Gill, and Vishu Kaur. It raised a $16 million Series A in August 2024 and a $60 million Series B in September 2025, as investors poured money into tools that promise to manage the code quality problems created by faster AI-generated software. (crunchbase.com) (coderabbit.ai) (techcrunch.com) The immediate test is whether teams trust an agent to remember the right things without exposing the wrong ones. CodeRabbit is betting that the next fight in developer tools is not only writing code faster, but keeping the rationale behind that code from disappearing into Slack. (docs.coderabbit.ai) (coderabbit.ai)