Issues

• Dependabot: the 12 filelock alerts wait on a LiteLLM upgrade past 1.105

  apache/tooling-llmao#77 · Filter by author gstein opened on Nov 13, 2026

• A failed save can mint a second key against one approval

  apache/tooling-llmao#76 · Filter by author andrewmusselman opened on Nov 12, 2026

• Inform harness about context size

  apache/tooling-llmao#75 · Filter by author mbien opened on Nov 10, 2026

• Let a request say "self-hosted only": tag the paid deployments, LiteLLM's tag filtering

  apache/tooling-llmao#73 · Filter by author gstein opened on Nov 5, 2026
  1

  

• Request priority on our own boxes: let administrative work yield to users

  apache/tooling-llmao#72 · Filter by author gstein opened on Nov 5, 2026

• Capacity recommendations: the remaining rules, a minimum rental count per model, Snooze, and a parked LLM analysis

  apache/tooling-llmao#70 · Filter by author gstein opened on Nov 3, 2026

• A scanner is walking model names as bearer tokens against /v1/models, daily since 2026-10-28; surface refused-at-the-door counts to admins

  apache/tooling-llmao#68 · Filter by author gstein opened on Nov 2, 2026

• Evaluate Laya as a self-hosted JEV-style classifier for dynamic routing

  apache/tooling-llmao#67 · Filter by author ANierbeck opened on Oct 30, 2026

• Upgrade LiteLLM to ≥ 1.106.x to enable open JEV-style classification (oss_classifier)

  apache/tooling-llmao#66 · Filter by author ANierbeck opened on Oct 28, 2026

• Dynamic model routing via LiteLLM's complexity router (config-only, opt-in)

  apache/tooling-llmao#65 · Filter by author ANierbeck opened on Oct 28, 2026

• Tracking: Add a real-time inference performance dashboard

  apache/tooling-llmao#64 · Filter by author PragmaTwice opened on Oct 26, 2026

• Match quota and budget displays to their configured periods

  apache/tooling-llmao#63 · Filter by author PragmaTwice opened on Oct 26, 2026

• Apply tier limit changes to existing API keys

  apache/tooling-llmao#62 · Filter by author PragmaTwice opened on Oct 26, 2026

• Enforce personal and project allowances in the inference request path

  apache/tooling-llmao#61 · Filter by author PragmaTwice opened on Oct 26, 2026

• Record request-level usage and distinguish quota charges from actual costs

  apache/tooling-llmao#60 · Filter by author PragmaTwice opened on Oct 26, 2026

• Replace the flat $0.32 per million tokens estimate for self-hosted models

  apache/tooling-llmao#59 · Filter by author PragmaTwice opened on Oct 26, 2026
  1

• Distinguish cached and uncached tokens in usage and quota accounting

  apache/tooling-llmao#58 · Filter by author PragmaTwice opened on Oct 26, 2026

• Use model-specific costs for quotas instead of raw token totals

  apache/tooling-llmao#57 · Filter by author PragmaTwice opened on Oct 26, 2026

• Tracking: improve quota accounting, cost estimates, and enforcement

  apache/tooling-llmao#56 · Filter by author PragmaTwice opened on Oct 26, 2026

• Qwen Code setup instructions

  apache/tooling-llmao#55 · Filter by author pjfanning opened on Oct 25, 2026
  1

• Secrets need to be redacted

  apache/tooling-llmao#54 · Filter by author andrewmusselman opened on Oct 22, 2026

• Model names and caps: where things stand

  apache/tooling-llmao#52 · Filter by author gstein opened on Oct 13, 2026
  

• investigate HF_TOKEN

  apache/tooling-llmao#50 · Filter by author gstein opened on Oct 11, 2026

• design discussion for /models

  apache/tooling-llmao#46 · Filter by author gstein opened on Oct 7, 2026
  2

• feature request: page data for configuring coding assistants

  apache/tooling-llmao#44 · Filter by author gstein opened on Oct 4, 2026