Genesis v1.19.92 Release Notes
September 10, 2026Highlights
- Pulse Mission Value now compares completed mission work against manual baselines, showing time savings, timing quality, coverage, and mission drill-down.
- Genesis Prover measures two database estates, accounts for every difference between them, and produces a verdict with evidence — now visible in every deployment’s sidebar.
- Genesis can now connect to a Palantir Foundry tenant to read datasets, lineage, pipelines, ontology, and Workshop apps, and run SQL queries.
Deployment Artifacts
Use these exact references when deploying this release:- Genesis release:
v1.19.92(promoted fromv1.19.92-rc.9) - Docker Hub image:
genesiscomputingai/genesis:v1.19.92 - Docker Hub digest:
genesiscomputingai/genesis@sha256:131dc1176de9479b9d502f28510f28e347c3907aea4aea704ebc0f47380c6ea0 - Helm chart:
oci://registry-1.docker.io/genesiscomputingai/genesis - Helm chart version:
0.3.89(appVersion: v1.19.92)
Upgrade Notes
- Installs that used the removed Mistral or Cerebras LiteLLM providers must select another AI provider after upgrading before chat works again.
- Saved embedding provider and model settings now always override environment variables; the LITELLM_EMBEDDING_MODEL variable is removed—use EMBEDDING_MODEL instead.
- The LLM Overrides page, its legacy APIs, and the FORCE_ENV_LLM_PROVIDER variables are removed; configure your provider and model through AI Providers.
Improvements
- Context Graph can now scan a chosen git branch of a repository instead of whichever branch the clone happened to check out.
- MCP connections can now send custom headers, connect over SSE, and be filled in from a pasted .mcp.json server entry.
Bug Fixes
- Context Graph lineage tools no longer disappear after startup, and chats through an OAuth-protected LiteLLM proxy no longer fail with missing bearer token errors.