Forage

This updates inference profiles so the upstream model field accepts the provider-native model identifier, for example accounts/fireworks/models/minimax-m3 for Fireworks or `gpt-4…

The GitHub issue classification worker now treats model-provider account and quota failures as delayed retries instead of failed jobs. It sanitizes provider request and response er…

Hive now configures Oban to use a local database peer implementation that preserves the peer loop when a transient database connection failure exhausts the election transaction. Th…

This adds Oban.Plugins.Lifeline to the application Oban configuration with a 30 minute rescue window.

The model gateway is now mounted at /inference/v1 instead of the root /v1 path. The inference profile detail page now shows that namespaced base address in the client configura…

create_spec (via MCP and the dashboard) could hang indefinitely — validation passed, then the create wedged server-side and MCP clients timed out. This bounds every spec write so…

create_spec fails in production for any summary longer than 255 characters. The failure is instant (~9ms) and surfaces to MCP clients as *“the connector’s server isn’t responding…

Raises the spec body character limit from 20,000 to 100,000.

• Start Oban unconditionally in the application supervision tree before the endpoint.

Hive’s model gateway now accepts embedding requests through POST /inference/v1/embeddings.