We're still building things here! You might notice a few rough edges while we work on improvements. Help us improve by reporting bugs here.

Organization Features

Explore the Companin organization platform: assistant lifecycle management, unified knowledge operations, session orchestration, conversation feedback, API-first controls, and observability built for teams.

What Your Organization Gets Out of the Box

Companin ships with the infrastructure your operations team needs to launch and supervise customer facing AI assistants. Every feature is backed by real API endpoints, rate limited for safety, and isolated per organization to keep data boundaries clear.

Assistant Lifecycle Management

Manage every assistant persona through the POST /api/v1/assistants/, PATCH /api/v1/assistants/:assistant_id, and DELETE /api/v1/assistants/:assistant_id endpoints. Store tone, language, and default task templates per assistant so each deployment reflects the audience it serves.

Version safe updates let you promote changes without touching previous configurations, while per assistant activation flags make it simple to pause experiments without reissuing API keys.

Unified Knowledge Operations

Upload PDFs, CSVs, plain text, JSON, or images with POST /api/v1/knowledge/files/, curate Q&A banks through /api/v1/knowledge/qas/, and ingest live URLs via /api/v1/knowledge/urls/. Each record captures AI generated summaries, keywords, and embeddings so your retrieval stack is ready the moment content lands.

Link or exclude sources per assistant to keep regulated material separated from public responses and refresh embeddings whenever your knowledge evolves.

Session Orchestration & Context

Use POST /api/v1/sessions/ to spin up authenticated or anonymous sessions, attach visitor traits through /api/v1/contexts, and store metadata that maps back to your CRM. Configurable expiration keeps stale sessions from lingering while still providing full traceability across channels.

Because context, sessions, and conversations share organization scoped identifiers, you can stitch together omnichannel journeys without leaking data between clients.

Conversation & Feedback Signals

The conversations API keeps every exchange auditable with POST /api/v1/conversations/ and GET /api/v1/conversations/:conversation_id/messages. Capture satisfaction scores using POST /api/v1/conversations/:conversation_id/feedback so product teams can measure outcomes without building additional plumbing.

Feedback records include rating, free form comments, and actor metadata, giving you a clean signal to improve prompts, escalation rules, or human handoffs.

API-First Governance

Every call requires an X-API-Key header that maps to a specific organization. The built in rate limiter caps traffic at 100 requests per hour per organization and records breach attempts for audit trails. Because we authenticate through Django Ninja, you can extend authentication with your own middleware when you need more complex policies.

Structured JSON responses keep consumers predictable across 2xx and 4xx states, making observability tooling easier to wire up.

Analytics & Observability

Dedicated analytics endpoints such as GET /analytics/logs/, /analytics/usage/, /analytics/endpoints/, and /analytics/rate-limits/ surface request level detail, aggregate volume, endpoint mix, and throttling events. Filter by organization or API key to spot anomalies before they hit SLAs.

Combined with conversation feedback and metadata, these dashboards give customer teams a full closed loop on performance without needing a separate data pipeline.