Archilas: Shared decision context for AI

Archilas is a context service for teams that want an AI assistant to refer to decisions made across multiple tools. It syncs connected sources and returns relevant dated notes and their origins to a compatible assistant through the Model Context Protocol.

How does it keep decisions current?

Archilas reads what the connected account can see, imports history and continues syncing. Its documentation says a newer decision can replace an older one without erasing the old note. Both dates remain, and a response should prefer the current note. That lets someone review where a decision came from and compare it with the note it replaced.

The service lists sources such as Slack, GitHub, Gmail, Notion and Linear. What a connection can see depends on the account and integration. Before linking a work tool, review what the connected account is allowed to read.

How does an assistant receive the notes?

The publisher documents a remote connection that uses OAuth. A compatible client can request team context, and Archilas says the result includes sources. Its documentation tells the client to say it does not know when the service returns no matching memory. The terms also warn that delivered context can be incomplete or outdated. The source is still useful for checking whether a quoted decision is current.

What does the free tier cover?

The free tier includes integrations and 100 metered context deliveries per month, which the product calls questions. Higher tiers have larger allowances. Archilas says paid arrangements are handled by email while its online checkout is closed. Teams comparing plans should estimate the number of context deliveries they expect and check which data will go to a connected assistant before adding broad workspaces.