Launching today. Tune in youtube.com/watch?v=3xd51mUi…

We're live!! youtube.com/watch?v=3xd51mUi…

The Future of CI xitter.cf/i/broadcasts/1aKbdEaYw…

Tomorrow: we're fixing CI for the agent era. Live at 10 a.m. PT / 1 p.m. ET. Grab the link to catch it live: nx.dev/events/2026-oct-produ…

Agents mean more code, more PRs, more pipeline runs. Oct 22, live: → Nx: updates so every run does less work → Nx Cloud: next-gen caching and distribution. Faster CI, less compute. All available the same day. Save your spot: nx.dev/events/2026-oct-produ…

One more thing: on Oct 22nd we're hosting a product event with a big announcement about scaling CI for the flood of PRs and tests your agents produce. Register here 👇 nx.dev/events/2026-oct-produ…

Full video walkthrough here: youtu.be/bNQ1svSw8VY

Nx 24 will remove the built-in executors that have an inferred task replacement (37 executors across 15 packages). The 23.x releases are a good time to migrate. We wrote up the whole story, including how to migrate 👇 nx.dev/blog/look-mum-no-exec…

Project graph creation got up to 60% faster on cold runs and up to 80% faster on warm runs in a large enterprise workspace. How much you gain depends on your repo and the machine it runs on.

When an AI agent runs Nx, it now gets the new summary output style by default. Nx prints the task counts and, for each failed task, the path to its full log. The agent reads a log file only when it needs the details, which keeps the token usage low.

The Oxlint plugin we shipped in 23.2 now runs in batch mode by default. Nx starts a single Oxlint process for all projects, and each project still keeps its own result, logs and cache entry. On a workspace with 50 libraries, linting went from 16.5s to 3.0s.

Task-based affected is opt-in in 23.3 and we plan to make it the default in Nx 24: NX_LEGACY_AFFECTED=false nx affected -t test If a task shows up (or doesn't) and you're not sure why, add --explain. It lists the affected tasks and exits without running anything.

Until now, nx affected worked at the project level. Change one spec file in a shared library and every project depending on it is affected too. 23.3 selects tasks by what they read, so that change only runs the test task of the project the spec file lives in.

🚀 Nx 23.3 is out. - Task-based nx affected - Agent-friendly task output - Batched Oxlint runs - Faster project graph creation - Cypress 16, Vitest 5, pnpm 12 And more. All the details 👇 nx.dev/blog/nx-23-3-release

CI was built for humans. Agents changed that. On Oct 22, we're showing how to make CI keep up without burning more compute. Register here: nx.dev/events/2026-oct-produ…

Nx v24 lands in a few weeks and drops the built-in executors, now replaced by inferred targets. Inferred targets use your tool’s own config to get runnable targets and cache inputs/outputs. @meeroslav got you covered with all the details: nx.dev/blog/look-mum-no-exec…

A 3-person platform team runs CI for a 600+ project Nx monorepo. With Nx Agents on Nx Cloud they: - maintain half the CI config and don't re-shard as new projects land - trust the cache again, thanks to task sandboxing How @WeAreLegora got there: nx.dev/blog/legora-success-s…

You can add Nx to an existing pnpm workspace without restructuring anything. Install nx at the root, add a small nx.json with your targetDefaults, and run tasks across every project with nx run-many. Caching and task ordering come from there.