With Rote set up, you talk to your agent as usual. Priya asks Claude Code for a ship report, gives feedback, and gets a result. Behind the scenes, Rote remembers every step, her feedback, and a rate-limit error. Rote automatically turns the successful path, a service-and-time check, and retry handling into a repeatable, shareable Play. Jamie asks Codex to run Priya’s report with new dates. The agent recalls the Play by its URI. Both people stay in their normal conversation with their agent.
The same job.
The same explanation.
Again.
You guide an agent through the logs, checks, and decisions. The job gets done, but the method stays in the conversation.
The next time this job comes up, another team member needs you to explain it again.
- You work out the methodThe logs to read. The checks to make.
- Keep the steps as a PlayYour method, ready to share.
- The next team member runs itThe same method. Their own inputs.
Find a Play for your next job.
Describe what you need done. Find Plays that fit, with a short explanation of why.
Starting your search
- Understand the job
- Search Plays
- Select matches
Search Community Plays. No installation needed to explore.
See what repeat work could save.
Keep the method you worked out. See how model costs could change when your team uses it again.
450,000 tokens avoided
How this adds up
45,000 saved per repeat run × 10 repeat runs = 450,000 tokens saved
Per repeat run 50,000 tokens finding steps − 5,000 tokens reusing steps = 45,000 tokens saved.
In dollars 450,000 ÷ 1,000,000 × $5.00 = $2.25.
Example rate: $5 per million input tokens. These are sample values, not measured results or a bill. Both paths share an initial 50,000-token run ($0.25), excluded from the repeat-run comparison. Output tokens, tool fees, and Modiqo plans are excluded. Actual costs depend on the Play, model, and work performed.
Measure the return on your team’s work.
Compare model cost per completed job, minutes spent supervising or retrying, and successful runs out of all attempts. Use measured results to estimate monthly savings at your team’s volume.
See the steps before you run them.
- Know the author. Follow the method back to the person and run it came from.
- Inspect the steps. Review the source, inputs, and checks.
rote play inspectshows what the Play reads and writes. - Choose how to run it. Use a supported agent or run through Rote directly. Check the Play’s tools and access requirements first.
01Who built it#!/usr/bin/env -S rote play run /** * @rote-frontmatter * --- * name: ship-report * provenance: * author: priya@modiqo.ai * workspace: ship-report # from the run's trace02Steps Rote runs* steps: * alerts: # trace 2 · kept * endpoint: adapter/datadog * method: monitors.search * params: { status: "triggered", since: $since } * prs: # trace 3 · kept * endpoint: adapter/github * method: pulls.list * depends_on: [alerts] * params: { since: $since, state: merged, touching: @alerts{.services} } * # trace 1 · read every commit since monday · dropped */03The result you getconst sdk = await import("__ROTE_PRESENTATION_SDK__"); const { FlowOutput, loadPresentationContext, stepName } = sdk; const out = new FlowOutput(), ctx = await loadPresentationContext(); const alerts = ctx.step(stepName("alerts")).body; const prs = ctx.requireAvailable(stepName("prs")).body; out.human(`${prs.length} shipped, ${alerts.length} broke, since ${ctx.params.since}`);
Make your expertise useful to someone else.
Turn a job you know well into a Play. Share it with your team or publish it for the community.
≡×CODE >walk me through this PR stack before I review it
Builds the whole pull-request stack from the bottom PR and explains each one in order.
≡×OPS >is anything we depend on down right now?
Checks the live status of GitHub, npm and Cloudflare from their public status pages.
≡×SALES · SUPPORT >what in my inbox needs a reply today?
Read-only Gmail triage. Sorts a window of mail, from headers alone, into what needs you.
Start with a problem you know.
A developer learning Rote turned six adoption questions into six Plays. PlayFit helps people find a Play for their work.
Explore PlayFit by jvxplaymaker ↗Jayanth KoppalaFounding Engineer at Qatobit A Play that reads the day’s coding sessions and commits and writes the post for me. Claims only.
Rana Singh@ranadeepsingh I reach for skills to guide different tasks, and Plays to rerun a method with new inputs.
Manan PaliwalDevOps engineer I started … planning to build one Rote Play. Ended up building four.
Harsha Priya GanapathyFull-stack & AI/ML developer Rote makes the workflow inspectable, repeatable and recoverable. A failed stage was resumed while four completed stages were restored instead of rerun.
Vedang P.Math & CS student, MIT Manipal Honestly, I haven’t had this much fun building in a while.
Jay@JayBosshq I steered it live, kept my wrong turn in, and it comes back every evening at 8 on its own.
Iniyan SFreelance developer What if a production debugging process could be captured and reused as an executable workflow?
PUBLIC POSTS · SEPTEMBER 2026
Put your team’s expertise within reach.
Share Plays in an organization so teammates can use the methods you’ve worked out together.
Create your organization →Bring us a job your team keeps repeating.
We’ll work through where a Play could help and what your team needs to use it.
Talk through your use case →Put a Play to work.
Find reusable steps for a job you need done. Or share a method you already know.
Ready to run or create? Set up Rote →
- 01INSTALLInstalls Rote, and Play, the sidekick that lets your agent drive it. Detects your harness, starts guided setup. Free for individuals. Keys never leave your machine.
curl -fsSL https://getrote.dev/playoffs/install.sh | sh - 02ASK$play hands the ask to the sidekick, and your agent does the job through Rote. Any harness: Claude Code, Codex, Cursor, Kimi, Hermes, or Pi.
$play what shipped this week, and what broke? - 03KEEPthe job worked. keep what made it work as a Play?▸ YES, SAVE THE PLAYPlay minted. A link, weekly, sent to #engineering.