Curated single-page applications built with React, Vite, and a shared state layer—powering experiences on trydojo.ai.
- Read the platform docs: docs.trydojo.ai
- Reinforcement-learning loops need safe, deterministic sandboxes; real production apps mutate too quickly. Our mock SPAs mirror the UX of live tools like Linear without exposing private data.
- Authentication is predictable here—agents can safely interact with environments without fretting over CAPTCHAs, SSO glitches, or expiring secrets.
- State-setting, verification, and retrieval stay under tight control: every view reads and writes through
dojo-hooks, so RL agents can rely on consistent schemas instead of chasing ever-changing production APIs.
@chakra-dev/dojo-hooksexposesuseDojoStateand a globaldojohelper for persisting and retrieving SPA state without wiring extra context.- Each app consumes these hooks to keep widgets in sync and to share state across routes or nested components.
- Duplicate an existing
*/appfolder (or scaffold a freshnpm create vite@latestReact app) inside this repo. - Install deps with
pnpm install(ornpm install) and add@chakra-dev/dojo-hooks. - Use
useDojoStateto hold the shared data you want to read/write across the experience. - Run
pnpm run devwhile building, andpnpm run buildto emit the static bundle that ships.
- Open the GitHub Issues tab and filter by the
bountylabel—each issue describes a SPA the team wants next. - Comment with your intent and wait for a maintainer to confirm assignment before you dive in.
- Ship the SPA in a PR that links the bounty issue; include any deployment notes.
- Once the PR merges, the maintainers close the issue with payout details so you can get paid for the contribution.