About Tracebrain

Most of what a company knows is said, not written down: in conversations with clients, in the team's own threads, in the promise made at the end of a call. When the person who heard it leaves or forgets, it is gone. Tracebrain reads those conversations and keeps what mattered in them: a profile for every person, a note for every day, the open promises as tasks and the work grouped by project.

Sources come in through connectors, and the connectors page lists which are live today and which come next. The memory goes out to any AI assistant through MCP, and it is plain Markdown that the company can download at any time. For companies says how a team works in one shared memory, and pricing what it costs.

Tracebrain started in August 2026 inside our own company. We were losing context: what a client asked for, what was agreed on a call, who promised what to whom. It lived in chats and conversations, never in a document, and when a person was busy, on holiday or gone, the knowledge went with them. A wiki did not help, because nobody had time to write things down.

So we built a company brain for ourselves: an AI knowledge base that fills itself from the conversations the team already has. Every night it writes down what was decided, promised and paid, keeps a profile of every client and colleague, turns promises into tasks and groups the work by project. Institutional knowledge stopped depending on who remembered it.

Then we noticed we were not the only ones. Other companies had the same problem and wanted the same thing: a second brain for their company, built from what it already says, that any AI assistant can read. So we decided to build Tracebrain for every company, not only ours.

The name comes from the memory trace, or engram: the record an experience leaves in the brain. Tracebrain is the trace a company's conversations leave. The product was called Trace at first and took the name of its domain on 1 October 2026. It is not trace.so, the Trace meeting app at traceapp.info or the tracebrain package on PyPI.

Write to [email protected]. Questions about privacy, deleting data, billing and partnerships go to the same address.