Product concepts and technical reference
This page explains the technical surfaces behind the product. Customer setup starts with the Unreal and AI assistant guides. SDK and CLI clients use these same routes and return structured envelopes with request IDs for recovery.
Repositories
Project name, status, branch, region, storage, bandwidth, health, and revision state.
Connection info
Remote URL, CLI commands, supported Lore version, and short-lived connection guidance.
Locks
Lock ownership, path, reason, timestamps, and status for binary asset coordination.
Transfers
Push, pull, clone, and sync status with byte totals, speed, client, and errors.
AI access
Read-oriented assistant access with repository limits, status, scopes, action counts, and one-time secret reveal.
Connect tokens
Temporary credentials for Lore connection setup only. They are not broad project mutation credentials.
AI assistant scopes
Assistant access should be explicit and narrow. Connection-token creation is not granted unless the scope says so.
| Scope | Meaning |
|---|---|
repo:read | Read repository metadata and status. |
activity:read | Read timeline events. |
locks:read | Read visible lock state. |
summary:create | Request OpenAI-backed activity summaries when configured. |
connect-token:create | Mint short-lived repo-scoped connect tokens; default-off for agents. |
Route groups
GET /v1/agent/sessionGET /v1/agent/manifestGET /v1/agent/environmentGET /v1/billing/catalogGET /v1/orgs/:orgId/billingPOST /v1/orgs/:orgId/billing/previewPATCH /v1/orgs/:orgId/billing/subscriptionGET /v1/reposGET /v1/repos/:repoIdGET /v1/repos/:repoId/connection-infoGET /v1/repos/:repoId/activityPOST /v1/repos/:repoId/activity/summaryGET /v1/repos/:repoId/locksPOST /v1/repos/:repoId/locksDELETE /v1/repos/:repoId/locks/:lockIdGET /v1/agent-tokensPOST /v1/agent-tokensDELETE /v1/agent-tokens/:tokenId