Miko
Miko is a self-hosted, open-source coding agent that connects to Slack, Linear, GitHub, and other channels to pick up assigned tasks. Every task runs in its own isolated Git worktree, keeping your code on your own machine.
Updated: 2026-10-27
What is Miko?
Miko is a self-hosted open-source coding agent under the Apache 2.0 license. Once connected to Slack, Linear, GitHub, GitLab, and Zulip, you can assign an issue or @Miko to get work started. It supports models including Claude Code, Codex, Cursor, Grok, Gemini, and OpenCode, and code runs on your own machine. It treats the agent as a member of the team, emphasizing integration with existing collaboration tools rather than building a separate workflow.
Who is it for?
- Technical teams that value control over their code and data and are willing to self-host.
- Engineering organizations already managing tasks with Slack or Linear that want agents to pick up work directly.
- Developers who want to switch freely between coding agent models.
Key features
5Self-hosted and open source
Apache 2.0 license, deployed in your own environment, so data and code never pass through a third-party platform.
Channel integrations
Connect to Slack, Linear, GitHub, GitLab, and Zulip to assign work directly within your everyday workflow.
Multi-model support
Officially supports Claude Code, Codex, Cursor, Grok, Gemini, and OpenCode, switching by task.
Isolated worktrees
Each task uses its own Git worktree, so tasks don't interfere with each other and can be reviewed separately before merging.
Local status board
View task status at 127.0.0.1:3456/board to see progress at a glance.
Things to know
Usage Tips
- Add labels to agent tasks to make it easy to track completion rates across different kinds of work.
- Start by assigning low-risk issues to Miko, and establish task templates and acceptance criteria.
- Use BYOK to configure your own model keys and assign models based on task characteristics.
- Track tasks with the status board, and check the changes in each worktree before merging. It suits teams that already have a well-defined development process; the agent is simply one more executor, while code ownership and review standards stay the same. Overall, how well this kind of tool works depends on whether the user's goals are clear. Before formal adoption, spend time reading the official documentation and changelog to confirm that the feature scope, data permissions, and billing rules match your business processes. When details are unclear, confirm them through official customer service or the help center first, and don't rely on third-party information for important decisions. At the same time, any automation should retain a step for human review; especially in scenarios involving money, external communications, or legal obligations, humans remain ultimately accountable. The value of a tool isn't in replacing judgment, but in handing off repetitive labor so people have time to tackle more complex problems.
Traffic & popularity
Data month Sep 2026
Miko receives about 2.3M visits per month.
Monthly visits
4 monthsEstimated from third-party traffic data.
Similar tools
LispaLispa is the cloud control plane for coding agents. It integrates Claude Code, Codex, Grok and Gemini into a single workspace. After the agent tasks are built and tested in the cloud, the results are viewed and merged through a browser.
TabbyOpen-source self-hosted AI coding assistant offering code completion, a question-answering engine, and in-line chat. The official states it supports deployment on consumer-grade GPUs, requires no external databases or cloud services, and keeps data within your own environment.
KodusOpen-source self-hostable AI code review tool, officially claiming that you can use your own model keys with no markup, unlimited PRs, the ability to identify issues across files and enforce team-specific custom rules.
Hermes AgentNous Research's open-source and self-hostable AI agent, officially available for free under the MIT license, supports persistent memory, automatic skill generation, multi-platform access, and isolated sandboxes. It can be downloaded for desktop or terminal installation.
WeWebWeWeb is an IT-approved AI application builder. Officially, it claims to be built with AI and edited with no-code, can be one-click published to a CDN or exported for self-hosting, emphasizes controllability and scalability, and is suitable for teams that need compliance.
AppFlowyFocused on a self-hosted open-source AI workspace, it integrates projects, Wikis, documents, and AI Q&A in one place. It supports local, cloud, and fully isolated environment deployments, allows the use of self-hosted large models, and the official states that the storage region can be specified by the user.