Career Ops UI

v1.105.0 Feature

This release adds 3 notable features for engineering teams evaluating rollout.

Published 21d Productivity & Wikis

View tool

✓ No known CVEs patched

Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

airbnb-style ashby ats career-ops claude-code cv

+11 more

express greenhouse-api hh-ru job-application job-search lever local-first no-telemetry resume self-hosted sse

Summary

AI summary

Adds an AI usage page showing token counts and estimated cost per provider.

Full changelog

[1.105.0] — 2026-07-06

AI usage & cost page. A new AI usage page (sidebar, next to Health) shows how many tokens you've spent on live AI generations — evaluations, reports, chats — broken down per provider over the last 24 hours, 7 days, 30 days, and all-time, with an estimated USD cost. Every live provider call appends a small {provider, in, out} record to data/llm-usage.jsonl (nothing is sent anywhere); runs with no API key (manual mode) cost nothing and aren't recorded.

New route module (30th) server/lib/routes/usage.mjs — GET /api/usage (read-only rollups) + server/lib/llm-usage.mjs (recordUsage normalizes the Anthropic/OpenAI/Gemini usage shapes and appends best-effort; readUsage/aggregate roll up per 24h/7d/30d/all window × provider) + server/lib/llm-pricing.mjs (an editable per-provider $/1M-token table — token counts are exact, dollars are approximate list prices you can correct to your plan; never billed). Recording is hooked at the dispatch chokepoints (runActiveProvider + routes/llm.mjs) so all live calls are captured.
New view public/js/views/usage.js (#/usage, window tabs). Tests: tests/usage-routes.test.mjs (normalize / price / record round-trip / aggregate / endpoint). 17 new i18n keys ×16 (usage.* + nav.usage). Help §6 extended in place.