SAGARIS Documentation

Build on the SAGARIS revenue engine

Guides, concepts, and API reference for the Company Brain, agents, and every channel.

QuickstartSet up a workspace and launch your first play.
Soon
Core conceptsThe Brain, agents, channels, and signals.
Soon
API referenceEndpoints, auth, webhooks, and SDKs.
ChangelogWhat shipped, every week.

The Company Brain

The Company Brain is the shared memory that every SAGARIS agent reads from and writes to. One record of truth across calls, email, SMS, and pipeline, so context never lives in a single channel again.

Every agent inherits the caller's permissions. The Brain resolves access at query time, so agents only ever see what the user is cleared to see.

How context is written

Activity is captured automatically as it happens. When a call ends, a message sends, or a meeting is booked, the event is normalized and written to the account record with its source attached, no manual entry required.

POST /v1/brain/context
curl https://api.sagaris.ai/v1/brain/context \  -H "Authorization: Bearer $SAGARIS_KEY" \  -d '{ "account_id": "acc_vertexa", "event": "call.ended" }'

Reading the Brain

Agents query the Brain the same way the API does. Ask for an account and you get the resolved record: the latest touch on every channel, the open threads, and the signals that changed the deal, already scoped to the caller's permissions.

GET /v1/brain/context/{account_id}
curl https://api.sagaris.ai/v1/brain/context/acc_vertexa \  -H "Authorization: Bearer $SAGARIS_KEY"

Permissions

Access is part of the memory, not a filter applied afterward. Every read resolves against the caller's clearance at query time, so an agent acting for a rep sees exactly what that rep would see, no more and no less. There is no separate access list to keep in sync.

Was this page helpful?

Book a demo