Federation

Shared knowledge. Clear boundaries.

A team may need the same product knowledge without sharing every conversation. Design that flow around dedicated Minds, named information scopes, and explicit access rules. Akasha provides the namespace and local replication foundations.

Start free Read the docs

Federation

Shared context.
Separate minds.

Bring the right knowledge across a boundary. Keep the rest where it belongs.

Product Mind

Selected for sharingRelease date · September 24 Product → Support

Internal planning notesPrivate

Only the selected context travels.

MINDS / FEDERATIONIllustrated example

Choose the context to share.

A product team shares its release date. Internal planning notes stay in its own Mind.

Choose a step to explore

Example sharing architecture. The daemon API provides local replication; cross-instance sharing is scoped with your deployment.

Plan the flow of knowledge

Decide what belongs where.

Separate shared reference material from each team’s working context. Define the source, destination, and access policy for each flow, with a deployment architecture suited to your regions and recovery needs.

Replication

Create another copy

A replication link copies data into a destination store. The current daemon interface uses a local destination directory.

Status

See whether the task is running

Inspect the replication task by its link identifier, and stop it when the copy is no longer needed.

Namespaces

Keep context scoped

Namespaces identify the owner, organization, agent, and space associated with a request.

Keyspaces

Organize before you share

Group records into named keyspaces so your application has a clear structure for the knowledge it stores.

Replication in the engine

Create a copy. Check its status.

Use the daemon API to replicate into a local destination directory and inspect the task’s running status. For connections across regions or instances, work with our team on the deployment architecture.

Create and inspect a local replication taskbash

# Run against your configured daemon with an authorized token.
curl "$AKASHA_URL/v1/replication/links" \
  -H "Authorization: Bearer $AKASHA_TOKEN" \
  -H "x-akasha-capability: $AKASHA_CAPABILITY" \
  -H "Content-Type: application/json" \
  -d '{"dest_data_dir":"/var/lib/akasha-replica"}'

# Use the link_id returned by the create request.
curl "$AKASHA_URL/v1/replication/status" \
  -H "Authorization: Bearer $AKASHA_TOKEN" \
  -H "x-akasha-capability: $AKASHA_CAPABILITY" \
  -H "Content-Type: application/json" \
  -d '{"id":"<link_id>"}'

Technical details

Define where the copy lives.

The daemon API manages local replication tasks. Cross-region destinations, recovery targets, and operating procedures are scoped with your deployment.

Local replication API

Create request: dest_data_dir · optional link_id
Destination: Local data directory
Status response: running: true or false
Create: POST /v1/replication/links
Stop: POST /v1/replication/stop
Status: POST /v1/replication/status

Namespace layers

Scope fields: tenant_id · reservoir_id · agent_id · dataspace_id
Access: Verified namespace scope
Dataspaces: Named spaces for stored knowledge
Keyspaces: Logical groups of stored records
Operations: Plan backup and restore separately
Recovery targets: Defined for each deployment

Keep exploring

Learning Isolation

Review incoming knowledge before it is accepted.

AkashaKV

The persistent storage engine behind Minds.

Memory

Organize the experiences and knowledge your agents use.

Plan the right flow for your knowledge.

Talk through your regions, data boundaries, and recovery needs with our team before choosing a federation architecture.

Discuss your deployment View pricing