v3.8.9 v3.8.10

Name: OmniRoute
Author: cheaper-inference

low risk

cheaper-inference/OmniRouteJun 4, 2026 → Jun 4, 20261 day between releases

What's New

Adds Provider Stats dashboard and OAuth quota resilience features

New Features

New Provider Stats dashboard with per-provider and per-model aggregates
Cross-request TTFT and latency tracking for observability
Token expiry visibility on provider cards
Podman/rootless deployment support
Responses stripping of previous_response_id for stateless clients

Improvements

Spaced quota sync for OAuth accounts to prevent bursting
Per-provider proactive-refresh skip list
Larger upload body-size limits for /v1/files
Sequential + spaced OAuth quota sync
Cross-request TTFT and gap-after-tool-call latency tracking

Bug Fixes

DeepSeek-web tool calls
Antigravity provider
Qoder provider
GitHub Copilot provider
Fireworks provider

Upgrade Notes

Upgrade adds a new Provider Stats dashboard and token expiry visibility. Configure the new 'previous_response_id' stripping setting for stateless clients like VS Code. Podman deployments now require the CONTAINER_HOST environment variable.

The Two Releases

v3.8.9

Jun 4, 2026· 5 months ago

This release adds an Obsidian context source with 24 MCP tools, Cursor vision input, and DeepSeek-web persistent sessions. It also improves Turkish locale-aware search and fixes SSE streaming for reasoning models and cache hits.

Full release notes

v3.8.10

Jun 4, 2026· 5 months ago

An OAuth resilience and observability release featuring spaced quota sync for OAuth accounts, a new Provider Stats dashboard, and Podman deployment support. It also adds token expiry visibility and fixes numerous provider issues like DeepSeek-web and Antigravity.

Full release notes