Table of Contents
Ditto is rolling out Ditto V2 through early access, introducing a shared Memory layer that connects AI agents with the conversations where project decisions often happen outside code repositories.
The company announced that Ditto V2 can connect selected conversations to one Memory and expose that context over the Model Context Protocol, or MCP. Ditto said each connection starts read-only, giving agents access to relevant discussion history without allowing them to write back into connected services by default.
The launch targets a common gap in AI-assisted software work. Coding agents can inspect code, issues and documentation, but they often miss decisions made in Telegram, Slack, Discord or other team channels.
Ditto's pitch is that an agent should be able to search those conversations alongside the technical artifacts it already tracks.
Ditto is "smart home" for agents, organizing conversations, files, projects, tasks and memory into context that can be reused across workflows. Its website says the product gives AI systems "a shared place to remember, retrieve, and act," with support for agents, files, memory, insights, live mode and connections across tools.
Antonio Verrico
Ditto V2 extends that positioning by focusing on work conversations that may never make it back into a repository or task tracker. Examples like a client changing requirements mid-build, a team debating a decision in Slack, or a bug being flagged in Discord before it is formally logged are things Dittos says never make it back to the repo.
More broadly, when teams are completing work, the authoritative record of a decision is often a chat thread rather than a pull request, ticket, or specification. That thread might be where the team agreed on a launch time, changed an implementation detail or clarified a customer requirement.
Ditto V2 makes that material searchable by agents when they are already working in development tools.
Get Early Access
