Project
Kotib Legal
kotib-legal
Overview
Platform for contract extraction and obligation management. It offers two extraction paths: a fixed workflow that pulls a predefined set of smart fields with simple three-state tracking, and an agent-based mode where the model analyses the document end to end, discovers every obligation rather than a fixed list, creates the contract record itself and reports progress through a nine-state workflow. Extracted obligations feed a tracker that fires renewal and deadline alerts. API responses are cached at four TTL tiers, throttled at 30 requests a minute for anonymous and 100 for authenticated users, with documents in object storage and metrics and dashboards wired in.
Key Results
Two extraction modes: fixed smart-field workflow and agent-based LLM discovery
Model-driven analysis that discovers every obligation, not a predefined list
Nine-state agent workflow with progress reporting
Async extraction with task-status polling endpoints
Obligation and deadline tracking with renewal alerts
Four-tier response caching (5 min to 24 h)
Per-user and anonymous rate limiting
Documents in object storage; metrics and dashboards for monitoring
Brute-force login protection