Skip to main content

Filters

September 26, 2026
Native Multi-Provider LLM & Fallback Failover

Fossa now provides first-class native support for all major AI providers directly through environment variables and in-app BYOK settings. You can seamlessly switch between Anthropic Claude (3.5 & 3.7 Sonnet), OpenAI (GPT-4o, o3-mini), Google Gemini (2.0 Flash & Pro), OpenRouter, and any custom OpenAI-compatible endpoint with zero proxy overhead.

We also introduced an automatic fallback failover engine. If your primary LLM encounters rate limits (429 HTTP), quota exhaustion, or temporary vendor outages, Fossa automatically falls back to your configured secondary provider without interrupting ongoing code reviews.

Per-Repository Model Routing

Organizations are no longer restricted to a single model family across all codebases. You can assign cost-effective, high-speed models for routine documentation and frontend repositories, while routing core security and architecture reviews to frontier reasoning models.

Stream Chunk Buffering: Fixed stream truncation on extremely large multi-file diffs when reviewing pull requests containing over 1,500 changed lines.

Reasoning Parameters: Normalized reasoning effort and temperature settings across Anthropic and Google Gemini APIs.

No changelog entries for this filter yet.

September 20, 2026
Hardened Production Self-Hosted Docker Stack

Self-hosting Fossa is now cleaner and more resilient than ever. The unified Docker Compose setup runs the NestJS REST API, asynchronous worker, webhook ingestion relay, and Next.js 15 dashboard with automated health checks, pre-configured database migrations, and RabbitMQ quorum queues.

The worker pipeline implements the transactional outbox relay pattern and idempotent inbox claims, guaranteeing that webhook events from GitHub, GitLab, Bitbucket, Azure Repos, and Forgejo are processed reliably even during network restarts.

Automated Schema Evolution & Health Checks

TypeORM PostgreSQL migrations and MongoDB audit history initializations now run automatically on container startup. Container health probes continuously monitor RabbitMQ queue backlog and database pool health.

Inbox Claim Locks: Resolved a lock contention edge-case where rapid consecutive git push events could delay job scheduling.

Connection Pooling: Optimized Postgres connection pooling under sustained concurrent webhook ingestion.

No changelog entries for this filter yet.

September 14, 2026
Fossy Rules Studio & Conversational Context Extraction

Fossy now proactively learns your engineering team’s conventions. When you explain an architectural rule, provide team style guidance, or clarify why a suggestion doesn’t apply inside a pull request comment, Fossy detects the convention and proposes saving it as an active Fossy Rule.

You maintain total control: Fossy displays the extracted rule for review and requests developer confirmation before persisting it to your organization or repository rule library.

Hierarchical Inheritance & Dry Run Previews

Define coding guidelines at the organization level that automatically cascade to all repositories, or override rules for specialized services using in-repo fossa-config.yml. With Dry Run mode, you can test rule changes against past pull requests to inspect suggestion precision before enabling rules in production.

Glob Path Matching: Fixed path glob matching for directory-scoped rules targeting deeply nested test files (**/*.spec.ts).

Inheritance Toggle: Added optimistic UI feedback when toggling organization rule inheritance with automatic rollback on network error.

No changelog entries for this filter yet.

September 8, 2026
Cockpit Engineering Hub & Shareable Review Views

Cockpit provides your engineering leads with real-time insight into active code reviews, suggestion acceptance rates, pull request turnaround time, and detected quality trends across all repositories.

All applied filters (time range, repositories, severity, and status) are synchronized directly to the browser URL, allowing engineers to copy and share exact filtered views with team members on Slack or Discord.

Granular Token Accounting & Spend Visibility

Full visibility into token consumption across models, teams, and repositories. Track prompt tokens, completion tokens, and estimated cost metrics in real time to prevent surprise cloud bills.

Timezone Graph Alignment: Fixed local timezone display calculation on Cockpit 30-day velocity charts.

RBAC Scoping: Enhanced team policy isolation to prevent unauthorized cross-repository telemetry reads.

No changelog entries for this filter yet.

September 1, 2026
Fossa 2.0 Open Source Platform Launch

We are thrilled to release Fossa 2.0 as a 100% open-source, community-driven AI code review platform. Built for engineering teams that demand complete data sovereignty, Fossa delivers enterprise-grade automated pull request reviews without locking your proprietary source code into third-party SaaS vendors.

Universal git provider integration covers GitHub, GitLab, Bitbucket, Azure Repos, and Forgejo out of the box with zero complex webhooks configuration.

Fossa CLI for Local Reviews & Pre-Commit Checks

Review uncommitted diffs and staged branches directly from your developer terminal using fossa review. Developers get instant architectural feedback before opening pull requests, reducing review turnaround latency and keeping CI pipelines green.

Initial Launch Optimization: Optimized container images, reduced bundle footprints, and added automated linting pipelines.

No changelog entries for this filter yet.