Different ergonomic from the Claude-led pair: instead of pasting individual files per turn, dump the entire repo into Gemini and let it reason across the whole graph. Claude takes over for tight, surgical edits; Gemini handles 'change auth across 14 files' moves.
2M+ token context lets you paste the entire repo (or 'git ls-files | xargs cat') as part of every prompt. Best when changes span many files.
When you know the file and the diff is local, Claude is faster and cheaper per turn. Use it for the last-mile edits.
Same role as the Claude-led stack. AI plugins draft from text; you refine.
Fast mode for Claude Opus 4.8 is being deprecated and will be removed on November 30, 2026. Developers using fast mode with Opus 4.8 must migrate to Claude Opus 5.5 to continue using the feature.
Anthropic released the Usage and Cost Analytics API in beta, enabling organizations to monitor token usage and costs across Claude API consumption with filtering and grouping by dimensions like model, workspace, and API key. The API provides /v1/organizations/analytics/usage and /v1/organizations/analytics/cost endpoints. Legacy usage reporting endpoints remain supported.
The maximum request size for the Messages API and Token Counting API doubled from 32 MB to 128 MB on the first-party Claude API, allowing developers to process significantly larger payloads in single requests. This change takes effect immediately and applies to both APIs.
The Claude Tag API is now available in beta for Claude Enterprise organizations, providing endpoints to read and configure Claude Tag settings on Slack channels and workspaces with Access bundle support. The API requires Admin API keys with read:claude_tag or write:claude_tag scope. This feature is exclusive to Claude Enterprise.
Figma + AI plugin: text to wireframe. Same as the Claude-led stack.
From your terminal: `git ls-files | xargs cat > /tmp/repo.txt`. Paste that as one block into Gemini.
I'm pasting my entire repo below as one block. Each file starts with a "// path/to/file.ts" comment line.
Repo:
"""
{{paste repo dump}}
"""
The change I need:
"""
{{describe change in plain English, e.g. "swap the in-memory session store for Postgres-backed sessions, keep the public auth API stable"}}
"""
Output, in this order:
1. **Affected files** — bullet list with a 1-line "why" each.
2. **Migration order** — which files to change first so the build never breaks.
3. **Per-file diff** — unified diff for each file, in migration order.
4. **One thing I should NOT change** — even if it looks tempting in this diff.
Constraints:
- Match existing style (imports, naming, error handling).
- No new dependencies unless explicitly necessary.
- If you spot a real bug while reading the repo, mention it at the END as "Side observation", do not silently fix it.Once Gemini gives you the cross-cutting plan, switch to Claude for individual file edits and bug fixes. Cheaper per turn, faster for local work.
Same as the recommended stack. Gemini's voice mode is fine here; the value is thinking out loud, not the LLM.
Deploy to Vercel. First sale before bedtime.
Swapped JWT-in-cookie for Postgres-backed sessions across 22 files in 90 minutes by pasting the whole repo into Gemini once and following its migration order. The Claude-led approach took 4 hours on a previous similar refactor because each file needed its own context paste.
`git ls-files | xargs cat` will happily include any .env files you committed. Audit before pasting; better, use `git ls-files -- ':!*.env*' ':!.git/*'`.
Gemini occasionally rewrites unrelated files when given the whole repo. Diff before you commit; reject any change you didn't ask for.
2M-token contexts get expensive fast on the API. Stay on the $20 chat tier for repos under ~150k LOC; switch to API only for bulk processing.
What changed in AI tools and the stacks that ship. One email a week.