OpenMayhem Studio
OpenMayhem Studio is the browser workspace for chatting with live models, building small software projects, and creating media with your OpenMayhem credit.
Browser-local projects
Studio stores project state in your current browser profile, not in a cloud project drive.
- Project metadata, conversations, and checkpoints use IndexedDB and are namespaced to your signed-in account.
- Project files use the browser's Origin Private File System (OPFS). Clearing site data, using a different browser profile, or losing the device can remove them.
- Export important projects as ZIP files; local storage is convenient, not a backup service.
Chat and project work
Studio opens in Chat for conversation, research, and coding. Keep the tab open while the model and browser tools work.
- Chat can read and edit project files, research the public web, and use browser tools with a compatible model. Type / to browse commands such as /build, /fix, /test, and /gsd:research; selecting a command fills the composer before you send it.
- Choose a live compatible model and optional routing constraints before a paid turn. Availability, provider count, context, pricing, and supported tools come from the live catalog.
- Set a maximum task budget and iteration limit. Studio stops scheduling new model turns when either limit is reached.
- Set the maximum output tokens and, when the selected model's signed contract permits it, its reasoning level. Studio shows only controls that model actually accepts.
- While a Chat turn is active, up to five more prompts can wait in order. You can remove a queued prompt before it runs.
- The context indicator estimates the percentage used and remaining. Auto-compact is on by default: older tool output is trimmed near 60% and summarized near 80%. You can turn it off; a turn that exceeds the model’s limit then pauses until you use /compact or shorten the context.
- Website credit is reserved before each model turn and settled from the provider's signed receipt. Unused reservation is released.
- The model can start up to two background subtasks at once, sharing the parent task’s budget and model. Expand their cards to inspect progress and results, or stop them individually.
- Questions appear inline in Chat with choices and a text field. Submit your answer to continue; answered questions remain in the conversation after reload.
- Attach up to three PNG, JPEG, or WebP images or MP4 or WebM videos in Chat when a compatible live provider is available. Reorder them before sending. Files are processed one at a time; later turns retain the processed answers without resending earlier file bytes.
- Scroll back through long conversations without losing your place as new output arrives. Use Jump to latest to return to the newest message.
Commands and background tasks
- Type / and use ↑/↓ to choose a command, then Enter or Tab to fill it in. Escape closes the menu. Outside the menu, ↑/↓ at the start or end of the input recalls your previous prompts and restores your unsent draft.
- /build [instructions] inspects the project, saves .planning/PLAN.md, changes actual project files and verifies the result after the last change. It records the checks in .planning/VERIFICATION.md. Missing evidence sends the model back to work within your task limits. /gsd:go uses this same build flow.
- /research [topic] searches from different angles, reads sources and produces a cited synthesis. Search snippets alone are not enough. The answer should separate supported findings, conclusions and gaps; completion checks do not guarantee that every claim is correct.
- /parallel [instructions] asks the model to split independent work between two background subtasks. For example: /parallel inspect accessibility and review the project tests. Each task has its own progress and result, shares the parent budget, and can be stopped separately. A third concurrent subtask is rejected.
- /tasks shows background tasks without starting a model request. Task cards appear once the model creates them; provider capacity can make them wait. Keep Studio open while they run. Model quality, your limits and available tools still affect whether a task succeeds.
Files, terminal, and preview
- Preview, Files & code, and Terminal are optional panels opened from Chat. The file tree and editor use the selected browser-local project; hiding a panel keeps its state.
- The terminal runs inside a WebContainer in the browser. It is not a shell on your computer or an OpenMayhem server.
- Chat opens alone by default. Open Preview to run generated index.html projects automatically and refresh after edits. Closing the panel keeps it closed. Use Start preview for other saved HTML projects; projects with dev/start scripts use their development server.
- Model-proposed shell commands, process input, and HTTP API requests require your review and explicit approval in Chat before they execute.
Review the exact command or request before approving it. Commands and packages can execute third-party code and send project data over the network.
ZIP import and export
- An imported archive may contain at most 2,000 files and must be at most 64 MiB compressed and 64 MiB after expansion. Each text file is limited to 2 MiB and each binary file to 16 MiB.
- Studio rejects absolute or traversing paths, rewritten unsafe paths, symbolic links, duplicate paths, and archives whose declared size changes during extraction.
Media
Use Media for live image, video, speech, transcription, music, and supported Comfy workflows. Chat can also present media choices inline or open the full Media form.
- Studio starts from the model's published sample request and signed request contract. It exposes common controls directly and puts less common admitted controls under More options.
- Generated request JSON is collapsed by default. Raw workflow JSON is not the default interface. Protected graph topology, model, VAE, LoRA, and part fields are not turned into arbitrary controls.
- Reference image, audio, or video inputs appear only when the selected route's contract permits them. File count, MIME type, individual size, and total size are checked before file bytes are read or encoded.
- The final request is revalidated against the same contract-aware media or workflow builder used by the paid Playground.
If a workflow has no approved sample-backed Studio configuration, Studio must say so rather than presenting an empty or unrestricted JSON editor.
What survives a disconnect
- Asynchronous media and workflow jobs run on the platform after submission. Media job history restores their status, receipts, and retained artifacts after navigation, reload, or reconnection.
- Generated outputs are private delivery-cache artifacts retained for at most 12 hours from job creation and are never backed up. Download anything you need before it expires.
- The browser coding loop, WebContainer terminal processes, and preview do not continue after the Studio tab or browser runtime closes. Reopen the project and start those processes again.
Browser requirements
- Use a current desktop browser that supports secure contexts, cross-origin isolation, SharedArrayBuffer, IndexedDB, and OPFS. Studio detects missing runtime capabilities and keeps terminal execution unavailable rather than weakening isolation.
- The Studio route uses dedicated cross-origin isolation headers required by WebContainer. Extensions, embedded browsers, enterprise policies, or reverse proxies that remove these headers can disable the terminal and preview.
Privacy and responsibility
- Files and project history remain in this browser unless you export or explicitly send their contents. They are not automatically synchronized between devices.
- Prompts, selected context, tool results, and permitted reference inputs sent for inference are processed by the selected OpenMayhem provider under that route's advertised trust tier. Do not assume standard providers offer confidential compute.
- Do not place secrets in prompts, project files, terminal commands, generated previews, or ZIP archives. Review model changes and command output before using generated code outside Studio.
Open Studio
Sign in, add credit if needed, then open OpenMayhem Studio. Use Chat for conversation, research, and coding, or Media for generation and workflows. OpenMayhem Studio