StackReplay week of Nov 9 NEW - Patch React Server Components RCE (CVE-2026-126187) #168
StackReplay week of Nov 9 NEW ← All apps 
StackReplay week of Nov 9 NEW StackReplay week of Nov 2 NEW StackReplay week of Oct 19 NEW StackReplay week of Oct 5 NEW StackReplay week of Sep 28 NEW StackReplay week of Sep 21 NEW
StackReplay
Your AI coding, measured. Scan your agent logs in the browser for tokens, models, speed and rhythm.
Changelog
StackReplay week of Nov 9 NEW StackReplay week of Nov 2 NEW - Pin catalog dates in tests past the review horizon #165
StackReplay week of Oct 19 NEW - Keep the sample recap inside the default period #163
StackReplay week of Oct 5 NEW - Format the linked history test so CI passes #160
- Scan history files in parallel #159
- Describe StackReplay as it is now #158
- Connect a linked Claude history with the chooser that follows links #157
- Explain how to connect a linked Claude history #156
- Give panels room so text never touches the edge #155
- Find every Claude Code history and never skip one silently #153
- Show a bar for every day with AI activity #154
- Fix what the second review found #152
- Fix the visual defects from the site review #151
- Make every number honest and every label clear #150
- Make shared cards unfurl on X and read clearly #149
- Add insight headlines to the overview #148
- Make share cards about volume and models, not speed #147
- Stop the folder chooser from hanging on a home folder #146
- Retry the GitHub calendar fetch once #145
- Rebuild the app as a terminal-style overview with speed and GitHub #144
- Refocus the app on the recap, stats and sharing #143
StackReplay week of Sep 28 NEW - Compute streaks over full history in local time #141
- Support multiple accounts per plan and restyle plan confirmation #142
- Redesign the recap with deep stats and accurate attribution #140
- Add the AI coding recap, share cards and new homepage #138
- Redesign the homepage around the AI coding recap #139
- Make featured homepage models easier to compare #136
- Bring mobile model search into view and explain benchmark gaps #135
- Target accessible comparison controls in navigation tests #134
- Clarify model pricing and correct sitemap freshness #133
- Add sourced RSS and JSON update subscriptions #132
- Add sourced AI update pages and shareable provider filters #131
- Resolve plan-detail Compare links before navigation #130
- Add Kimi benchmark evidence with explicit unknown publication dates #129
- Add sourced Qwen, GLM and MiniMax benchmark results #128
- Export readable benchmark comparison images #127
- Add sourced Epoch GPQA Diamond evidence #126
- Make benchmark evidence clearer and exportable #125
- Make mobile plan comparisons easier to read #123
- Add sourced API token cost estimates #124
- Put public model and provider discovery first #121
- Add a Devin Teams seat cost calculator #122
- Add sourced provider directory and hubs #119
- Add sourced Amazon, Cohere and Llama model coverage #120
- Add sourced Mistral Large 3 API pricing #118
- Add Devin and licensed Gemini Code Assist coverage #117
- Add missing Kiro tiers and Cursor Teams coverage #116
- Fix public page previews, freshness dates and benchmark empty state #114
- Catalog review: Google, Antigravity and Copilot, October 3 #115
- Update October catalog disclosures #113
- Consolidate docs and fix stale references #112
- Avoid per-shard apt setup with a pinned Playwright container #111
- Make catalog watcher signal-driven #109
- Repair catalog source references #108
- Reconcile release status and catalog watcher triage #107
- Name Claude accounts from their local profile, hidden until shown #105
- Avoid unnecessary cross-location signature reads during browser scans #104
- Make multiple accounts per provider first-class in workloads and My Stack #103
- Reuse saved workload pricing and show intentional loading states #102
- Recompose the homepage around a lead market story and personal relevance #41
- Price multi-tool workloads past the candidate search route bound #40
- Raise browser intake limits for large agent histories #39
- Lead with a sourced AI market feed and rebuild My Stack's first screen #38
- Reset homepage around public and personal AI intelligence #35
- Add verified Sol scores and benchmark row highlights #37
- Run the Playwright suite in four parallel shards #36
- Turn My Stack into a subscription decision surface #33
- Add sourced model benchmark comparisons #34
- Admit Gemini 4 Argon announcement facts #32
- Polish My Stack editing, model access, and recovery #31
- Add My Stack overview and multiple-plan confirmation #30
- Support OpenCode history in the browser alongside Command Code #29
- Add automatic stack discovery with explicit current-plan confirmation #28
- Make Compare rows parallel, rename ChatGPT Pro tiers and add use in other apps #24
- Separate saved Replay compatibility from catalog metadata #23
- Feature GPT-6.1 Sol on the Models page #22
- Classify catalog coverage and add release-date sorting #21
- Detect provider source changes and new model coverage #20
- Pin Suggested Replay to the admitted DevDay snapshot #19
- Make Workload to Replay decisions visible and consistent #18
- Allow Cloudflare Web Analytics under CSP #17
- Preserve and price Command Code DeepSeek V4.1 Flash Fast routes #16
- Version subscription terms and add OpenAI processing-tier replay #15
- Fix homepage hero sticking on Replaying #3
- Fix CodeQL findings in import paths and benchmark server #14
- Pin miniflare's undici to 7.29.1 #13
- Models, subscriptions and compare polish: promo prices, cleaner model pages and a model matrix in compare #12
- Models batch A: lineup links, 18 new models and a table view #11
- Add catalog coverage audit and models roadmap #10
- Complete published plan and model decision details #9
StackReplay week of Sep 21 NEW