A dedicated computer
A managed cloud computer for your agent, with its project context, connections, and work history. It keeps working when you close Slack or your AI assistant.
$149 a month per workspace.
Invite your whole team at no extra charge. Bring your own AI provider: OpenAI (including Codex), Anthropic, GitHub Copilot, xAI, OpenRouter, OpenCode Zen and Go, DeepSeek, Kimi For Coding, MiniMax, Z.AI, and more.
Every workspace includes $20 in cloud credits each month to run separate coding sandboxes and use managed web search. Top up when you need more.
What’s included.
Your whole team
Invite everyone who needs to work with your agent. Teammates can add context, follow investigations, and review fixes. There is no per-seat charge.
Slack and your AI tools
Give your agent its own name in Slack, or work with it through the AI assistants you already use. MCP connects them to the same project context and tools.
Shared knowledge and memory
Keep project knowledge, findings, decisions, and completed work available for the next issue. Your team can continue where the last investigation left off.
Separate coding sandboxes
Independent coding tasks get their own computer, branch, and checkout. Use the included cloud credits to run them, and top up when you need more.
Unlimited Loops and skills
Set up support work to run on a schedule or when something happens. Give your agent reusable skills and instructions for how your team works.
Connected tools
Connect repositories, documentation, support inboxes, issue trackers, and monitoring through MCP. Your agent can use them to investigate issues and carry out work.
Vault for credentials
Store secrets encrypted and reference them by name. Supply the credentials and environment variables a task needs without pasting them into conversations.
Frequently asked questions.
What does an AI support engineer do?
Maito investigates customer issues using the repositories, documentation, and tools you connect. It gathers evidence, explains findings, and can prepare a proposed code fix. Your team reviews the evidence and decides whether to merge a change or send a customer response. It can ask for missing information and does not guarantee that every issue can be resolved.
What should we connect before our first investigation?
Create a workspace, connect your AI provider, and add the relevant repository and project documentation. Connect Slack or your AI assistant through MCP, then enable the tools needed to inspect the issue, such as logs or a ticket system. Access depends on the permissions and credentials your team supplies. Start with a specific customer report and the checks you expect the agent to run.
Who approves a proposed fix?
Your team reviews the proposed change, its evidence, and the reported checks. Coding tasks run in separate sandboxes and return changes for review. Check that the fix addresses the customer issue and use your normal code-review and merge process. A demonstration of passing checks is not a promise that every future task will pass.
What kind of coding work can Maito help with?
Maito can help with the engineering work around your code: investigating bugs, implementing changes, reviewing code, running checks, updating documentation, and following up on recurring issues. Give your agent a goal and connect the repositories and tools it needs. It uses your project context to carry out the work and bring back findings or changes for your team to review.
Can I name my agent?
Yes. Choose a name for your agent when you create your workspace. Give it a role, personality, and instructions that fit your team, whether you want help with maintenance, code reviews, QA, or customer issues.
Can my agent review code?
Yes. Give it a pull request, a diff, or a connected repository and tell it what to review. Ask it to look for bugs, edge cases, missing tests, or changes that conflict with your team's conventions. It can explain its findings with references to the relevant code, and you can ask it to work on the fixes in the same conversation.
Can my agent do QA?
Yes. Describe the behavior you want to verify and give the agent access to the project and testing tools. It can run configured checks, reproduce reported problems, investigate failures, and report what passed or failed. With browser tools available in its environment, it can also check web workflows. Tell it which scenarios matter and ask for evidence of the result.
Can my agent implement features and refactor code?
Yes. Give it a feature request, issue, or refactoring goal, along with the expected behavior and constraints. It can inspect the repository, make changes, run the configured checks, and prepare a pull request. Your team reviews the changes and decides whether to merge them.
How do we work with Maito?
Connect your workspace to Slack or an AI assistant through MCP. Start an investigation, then keep working in the same conversation. Teammates can add context, ask questions, review the findings, and continue the work without starting over.
Can Maito fix bugs in our code?
Yes. Connect a repository and ask Maito to investigate or fix an issue. Independent coding tasks use separate sandboxes, branches, and checkouts. Maito runs the configured checks and returns a pull request for review. Follow-up requests continue the same task, and your team decides whether to merge the changes.
Can my agent write tests and update documentation?
Yes. Ask it to cover a regression, add tests for a workflow, or update documentation after a code change. Share your project's testing commands and documentation conventions so it can follow the same process as your team. It can run the available checks and include the results with its work.
Can we give our agent our own engineering process?
Yes. Add project knowledge, instructions, and reusable skills for how your team works. A skill can describe your code-review checklist, QA process, release checks, or approach to investigating bugs. Your agent can reuse those instructions in conversations and Loops, with the tools and permissions you provide.
Can we work on several coding tasks at once?
Yes. Independent coding tasks use separate sandboxes, branches, and checkouts so each change has its own workspace. Keep follow-up requests in the original conversation to continue that task. Your team can review the results separately and decide how to merge changes that touch the same code.
Does Maito keep working when we close Slack?
Yes. Your workspace agent runs on a dedicated cloud computer, so closing Slack or your AI assistant does not stop its work. If it needs more information, it asks in the conversation. You can return to the saved findings and continue when you’re ready.
Which AI providers can we use?
Maito supports OpenAI (including Codex), Anthropic, GitHub Copilot, xAI, OpenRouter, OpenCode Zen and Go, DeepSeek, Kimi For Coding, Moonshot AI, MiniMax, Z.AI and Z.AI Coding CN, Meta, Ant Ling, Groq, Cerebras, Together, Fireworks, Baseten, Hugging Face, NVIDIA, Amazon Bedrock, Azure OpenAI, Cloudflare Workers AI and AI Gateway, Vercel AI Gateway, Radius, TypeSafe, and Xiaomi Token Plan AMS, CN, and SGP. Choose a model for your workspace, and optionally a different provider and model for coding tasks.
Can I use an existing AI subscription or API key?
Authentication depends on the provider. OpenAI supports ChatGPT subscription sign-in and API keys; the legacy Codex connection uses ChatGPT sign-in. GitHub Copilot, xAI, Kimi For Coding, Meta, OpenRouter, and Radius also offer sign-in. Anthropic requires an API key with separate API billing; Claude subscription sign-in is not available. Other providers use their supported API keys, tokens, or cloud credentials. Your AI provider bills its usage separately from Maito.
What context and tools can we connect?
Connect your repositories, documentation, support inboxes, issue trackers, and monitoring tools. Maito uses MCP integrations to research issues and carry out work with the access you enable. Store required credentials in Vault and reference them by name instead of pasting secrets into conversations.
Does my agent remember previous work?
Maito saves conversations and learns useful facts and reusable procedures from completed shared work. Later tasks can build on those findings and team decisions. Private conversations stay private unless you explicitly ask to share a fact. You can ask the agent to remember, correct, or forget information.
Can my agent take care of ongoing engineering work?
Yes. Set up a Loop with a goal, schedule or trigger, and the tools it is allowed to use. For example, ask it to investigate new errors each morning, check unresolved bugs, or prepare a weekly engineering summary. Loops run on a schedule, from a webhook, or when you start them manually. Your team can inspect the results, answer questions, and pause or resume the work as needed.
What is included for $149 per month?
One workspace with a shared AI agent, a dedicated cloud computer, unlimited team members and Loops, shared knowledge and skills, Slack access, MCP connections, and Vault. There is no per-seat charge. Every workspace also includes $20 in cloud credits each month.
What are cloud credits for?
Cloud credits pay for managed services your team uses, such as web search during an investigation. The included $20 allowance renews each billing month, and unused included credits do not roll over. You can buy more when needed; purchased credits do not expire.