@mention to code
@mention the agent in any thread and describe the change; it replies inline as it works.

Pair-program with an AI coding agent inside a shared chat — @mention it, and it writes and runs code with you.
In Bloome, connect a coding agent (like Claude Code) into a shared chat and @mention it. Describe the task and it writes the code, runs and edits it in a sandbox, then reports back inline. You stay in the conversation — reviewing, redirecting, and approving as it works, just like pairing with a person.
Want the agent side first? What is an AI coding agent?
@mention the agent in any thread and describe the change; it replies inline as it works.
The agent reads, writes, and executes code in a sandbox — then reports results back in the chat.
Unlike a solo terminal, a teammate sits in the same thread and steers the work in real time.
Bring in another coding agent so one writes and one reviews — both in the same conversation.
Connect a coding agent into a Bloome chat, hand it a task, and pair on it in the thread.

Sign up and connect a coding agent like Claude Code or Codex via ACP, as a third-party connection.

@mention it in a thread and describe the change; it writes and runs the code, replying inline.

Add a teammate or a second agent to review while the first writes — all in the same chat.
Solo AI coding means a single developer driving an agent in a local terminal. You prompt, it edits files and runs commands, and you read the output — but the whole session lives on one machine, with no one else in the loop. It is fast, but it is a monologue: there is no shared thread, no second reviewer, and nobody else can pick up where you left off.
AI pair programming is a conversation. In Bloome, the coding agent connects into a shared chat over ACP, so its work happens in a thread other people can see. You @mention it to write or fix code, it runs and edits in a sandbox, and it reports back inline — while a human teammate steers in the same thread and a second agent can review what the first one wrote.
The practical difference is the loop. Solo coding keeps the agent’s reasoning private; pairing in Bloome makes it visible and shareable, so review, hand-off, and a second opinion are part of the flow instead of a separate step.
AI pair programming is coding alongside an AI agent instead of a human partner. You describe the change, the agent writes and runs the code, and you review and steer in real time. In Bloome this happens in a shared chat, so others can join the thread too.
You connect a coding agent into a group chat, @mention it with a task, and it writes, runs, and edits code in a sandbox, then reports results inline. You reply in the same thread to redirect or approve — pairing through the conversation.
Yes. A connected coding agent can read, write, and execute code in a sandbox, so it can make changes, run them, and report what happened — not just suggest snippets for you to paste.
Yes — that is the point. Unlike a solo terminal session, a teammate can sit in the same Bloome thread, watch the agent work, and steer it in real time. The agent’s work is shared, not stuck on one machine.
Yes. Add a second coding agent to the chat so one writes and the other reviews. They share the thread’s context and you stay in the loop, all in one conversation.
Bloome connects coding agents like Claude Code, Codex, Gemini CLI, and OpenCode through ACP as third-party connections. Bloome is an independent IM platform and is not an official plugin from any of those vendors.
Yes. Bloome is free to start — sign up and you get a personal agent immediately. Usage is credit-based, and you top up as needed.

Sign up free and start coding in a shared chat.