Available now
Nothing to do during a session. Watch every plan on a clean dashboard, with a Kanban board of its tasks.
One command opens it in your browser. Each project gets its timeline, its rules, a board for every plan an agent made, and a page to confirm what the agent wrote.
With Pro or a team seat, one switch on a project’s page sends its history to central.ambih.com and keeps it in step. Open it there from any browser: the same plans, boards and history.
Install once, run central init in a repository, and Claude Code records the work from then on. These are the commands you will actually use.
$ npm install -g @ambih/coding-agent-memory # the command line, the MCP server and the dashboard
$ central init # register this project and connect Claude Code to it$ central dashboard # open the dashboard in your browser$ central doctor # check this machine and Claude Code's setup here
$ central plan list # the plans agents made in this project$ central plan show <id> # one plan, its steps and where each task stands$ central later # things noticed that should be done later$ central context src/app.ts # what an agent is told before it edits that file
$ central review # what agents wrote that nobody confirmed yet$ central approve <id> # confirm it$ central reject <id> --reason "..." # reject it; the agent hears why on its next call
$ central login https://central.ambih.com # sign in with your Ambih account$ central sync --on # in a project: send its history and keep it in step$ central sync --status # what is waiting to be sent
$ central agent cursor # print the setup for Cursor, Codex or VS Code$ central --help # every command and its options
A commit says what changed. It rarely says why, what was tried first, or which rule it must not break. That is what goes missing when people leave, and when an agent writes the code.
Intents, decisions, rules, ideas and rollbacks are typed records, linked to each other and to the parts of the code they are about. Ask about one file and you get everything that applies to it.
The MCP server hands an agent the context for a file before it touches it: the rules that apply, the decisions behind it, the rollbacks that already happened. With Claude Code, Central also records the work in the background. The plan you accept, its tasks, the files each one changed and the bugs it noticed are written down without a single prompt. When a session starts, Claude already knows what is still open.
The command line, the MCP server and the dashboard all run locally, with no account. Free covers every project you have on one machine.
Pro keeps the same history on all of your own machines. Team shares a project with the people who work on it, each with their own role.
Code Helper shows what the system is and where things connect. Central shows when it changed and why.
Using this? Other developers would find your experience useful. Write a review
One command, then register a repository.
npm i -g @ambih/coding-agent-memory
While you are here.
Watch what your software actually costs to run: memory, CPU, and the moment a build got slower.
Windows will show an “Unknown publisher” warning the first time you run it. Choose More info, then Run anyway.
See how a codebase fits together: what each file does, what depends on what, and where it touches a database or an API.
Windows will show an “Unknown publisher” warning the first time you run it. Choose More info, then Run anyway.
Record your screen to a video file. No account, no watermark, no time limit.
