Kilo for Slack Brings AI Coding Agents Into Developer Chats

Kilo for Slack embeds an AI coding agent directly in Slack, letting teams query repos, debug, implement fixes, and open PRs without leaving chat. It uses MiniMax M2.1 by default (free first week) and supports multi-repo context and multi-turn threads.

•
Kilo Code
Kilo for Slack Brings AI Coding Agents Into Developer Chats

TL;DR

  • Embeds an AI coding agent into Slack to run repo queries, debug, implement fixes and open PRs; announced Jan 16, 2026.
  • Multi-repository inference, continuous thread context, and cross-surface execution (Slack → IDEs/cloud agents/CLI).
  • Codebase queries, on-the-fly debugging from pasted stack traces, convert thread conclusions into code+PRs, and mass edits/refactors.
  • When @Kilo is mentioned, reads thread, accesses connected GitHub repos, spins up a cloud agent, creates a branch, commits changes, and posts the PR back in Slack; integrates with VS Code, JetBrains, and CLI.
  • MiniMax M2.1 as default model (free first week); over 450 selectable models; usage-based pricing billed same as direct model use.
  • Quick setup: create a Kilo Code account, connect GitHub in Integrations at https://app.kilo.ai, add the Slack integration, then mention @Kilo to begin.

Kilo for Slack embeds an AI coding agent directly into Slack conversations, aiming to reduce context switching by letting engineering teams ask about repositories, debug issues, implement fixes, and push PRs without leaving chat. The integration was announced Jan 16, 2026 and ships with MiniMax M2.1 as the default model (free for the first week).

What Kilo for Slack does differently

Kilo frames the Slack integration around execution rather than single-turn interactions. Its headline distinctions are:

  • Multi-repository inference by default: the bot attempts to infer which repository a thread refers to rather than being tied to a single repo.
  • Continuous, context-aware conversations: threads and pull requests can carry multi-turn context so follow-up requests build on prior exchanges.
  • Cross-surface execution: handoffs are supported across Slack, IDEs, cloud agents, and the CLI so generated work can be pushed as branches and PRs.

These design choices are positioned against other Slackbots that limit scope to one repo or one-off interactions.

Key capabilities

Kilo for Slack supports four primary workflows directly inside Slack threads or DMs:

  • Codebase queries: ask how a subsystem implements error handling or other patterns.
  • On-the-fly debugging: paste a stack trace and request diagnosis and a PR with a fix.
  • Implement fixes from discussion: convert thread conclusions into code changes and PRs.
  • Mass edits and refactors: request repository-wide changes (for example, updating a string across files) and have the bot create a branch and PR.

How it works

When @Kilo is mentioned in a thread, the bot reads the thread context and accesses connected GitHub repositories. For implementation requests, it spins up a cloud agent, creates a branch, commits changes, and opens a PR that appears back in Slack. Integration with existing Kilo tooling (VS Code, JetBrains, CLI) is intended to keep generated work compatible with established developer workflows.

Models and pricing

Kilo for Slack is model-agnostic in practice:

  • MiniMax M2.1 is the default model for the Slack integration and is free to use in Kilo for Slack for the first week.
  • Over 450 models are available to select from for the bot.
  • Pricing is usage-based and billed the same as using a model directly through Kilo.

Kilo positions the MiniMax partnership in the context of increased capability from open-weight models in 2025; the announcement links to broader reporting on model convergence and MiniMax’s recent Hong Kong IPO.

Getting started

Setup is described as a quick process:

  1. Create a Kilo Code account.
  2. Connect GitHub repositories in the Integrations tab at app.kilo.ai.
  3. Add the Slack integration from the same Integrations page.
  4. Mention @Kilo in any channel or DM to begin interactions.

Where this fits

Bringing an agent into Slack places code-focused automation at the point where teams already discuss issues and decisions. By retaining thread context and producing PRs without explicit context transfer, the integration aims to reduce repetitive explanation and friction between discussion and implementation.

Read the original announcement: https://blog.kilo.ai/p/announcing-kilo-for-slack

Continue the conversation on Slack

Did this article spark your interest? Join our community of experts and enthusiasts to dive deeper, ask questions, and share your ideas.

Join our community

Related Articles

GPT-5.2 Leads in AI Code-Review Test of Three Models

GPT-5.2 Leads in AI Code-Review Test of Three Models

Kilo benchmarked GPT-5.2, Claude Opus 4.5, and Gemini 3 Pro on a 560-line TypeScript PR seeded with 18 issues. GPT-5.2 found the most and flagged an auth bypass; Claude was fastest with perfect security detection; Gemini exposed N+1 and CSV bugs but missed an admin check.

1 shared tag
MiniMax-M2.1 Goes Free on Kilo: 230B MoE Coding Model

MiniMax-M2.1 Goes Free on Kilo: 230B MoE Coding Model

MiniMax-M2.1, a 230B-parameter Mixture-of-Experts coding model (10B active at inference) with a massive context window, is available free in Kilo Code for a limited time. Early benchmarks and Kilo testing show competitive performance versus top proprietary models.

1 shared tag
Kilo Launches Instant AI-Powered Code Reviews for GitHub PRs

Kilo Launches Instant AI-Powered Code Reviews for GitHub PRs

Kilo's Code Reviews runs automatically on GitHub PRs, posting inline comments and a summary the moment a PR is opened or updated. Teams can set focus areas and choose Strict, Balanced, or Lenient review styles.

1 shared tag