[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-11592":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":17,"stars7d":18,"stars30d":19,"stars90d":16,"forks30d":16,"starsTrendScore":20,"compositeScore":21,"rankGlobal":10,"rankLanguage":10,"license":22,"archived":23,"fork":23,"defaultBranch":24,"hasWiki":23,"hasPages":23,"topics":25,"createdAt":10,"pushedAt":10,"updatedAt":26,"readmeContent":27,"aiSummary":28,"trendingCount":16,"starSnapshotCount":16,"syncStatus":29,"lastSyncTime":30,"discoverSource":31},11592,"autopreso","kunchenguid\u002Fautopreso","kunchenguid","Realtime speech to presentation. Let the whiteboard whiteboard itself.","",null,"JavaScript",373,50,31,1,0,5,8,90,15,5.12,"MIT License",false,"main",[],"2026-06-12 02:02:32","\u003Ch1 align=\"center\">autopreso\u003C\u002Fh1>\n\n\u003Cp align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fkunchenguid\u002Fautopreso\u002Factions\u002Fworkflows\u002Fci.yml\">\u003Cimg alt=\"CI\" src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002Fkunchenguid\u002Fautopreso\u002Fci.yml?style=flat-square&label=ci\" \u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fkunchenguid\u002Fautopreso\u002Factions\u002Fworkflows\u002Frelease-please.yml\">\u003Cimg alt=\"Release\" src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002Fkunchenguid\u002Fautopreso\u002Frelease-please.yml?style=flat-square&label=release\" \u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fwww.npmjs.com\u002Fpackage\u002Fautopreso\">\u003Cimg alt=\"npm\" src=\"https:\u002F\u002Fimg.shields.io\u002Fnpm\u002Fv\u002Fautopreso?style=flat-square\" \u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fplatform-macOS-blue?style=flat-square\">\u003Cimg alt=\"Platform\" src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fplatform-macOS-blue?style=flat-square\" \u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fx.com\u002Fkunchenguid\">\u003Cimg alt=\"X\" src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FX-@kunchenguid-black?style=flat-square\" \u002F>\u003C\u002Fa>\n  \u003Ca href=\"https:\u002F\u002Fdiscord.gg\u002FWsy2NpnZDu\">\u003Cimg alt=\"Discord\" src=\"https:\u002F\u002Fimg.shields.io\u002Fdiscord\u002F1439901831038763092?style=flat-square&label=discord\" \u002F>\u003C\u002Fa>\n\u003C\u002Fp>\n\n\u003Ch3 align=\"center\">Let the whiteboard whiteboard itself.\u003C\u002Fh3>\n\n\u003Cp align=\"center\">\n  \u003Cimg src=\"https:\u002F\u002Fraw.githubusercontent.com\u002Fkunchenguid\u002Fautopreso\u002Fmain\u002Fassets\u002Fautopreso.png\" alt=\"autopreso whiteboard hero screenshot\" width=\"960\" \u002F>\n\u003C\u002Fp>\n\n> [!WARNING]\n> autopreso is in **alpha** and under active development. Expect rough edges, breaking changes, and the occasional weird drawing. Bug reports welcome.\n\nYou wanted to give the talk, not build the deck.\n\nautopreso runs a local web app with a live Excalidraw canvas and a listening agent.\nYou speak; transcripts stream to a model; the model draws, labels, and rearranges the whiteboard in real time.\nStage a few seed elements, hit start, and present.\n\n- **Hands free** - your speech drives an agent that edits an Excalidraw scene as you talk, no clicking required.\n- **Bring your own model** - use your OpenAI API key or Codex subscription. Auto Preso itself is completely free and open source.\n- **Can run locally** - use Moonshine for transcription and Ollama for the agent and you get a fully local setup.\n\n## Quick Start\n\n```sh\n$ npx autopreso              # boots the server, opens the browser\nautopreso listening at http:\u002F\u002F127.0.0.1:3210\n\n# In the browser:\n# 1. Drop reference materials onto the staging canvas (title, agenda, etc).\n# 2. Pick your microphone, transcription model, agent model, and optional Agent instructions.\n# 3. Click \"Start Preso\" and start talking.\n```\n\n## Install\n\n**npm (recommended)**\n\n```sh\nnpm install -g autopreso\nautopreso\n```\n\n**npx (no install)**\n\n```sh\nnpx autopreso\n```\n\n**From source**\n\n```sh\ngit clone https:\u002F\u002Fgithub.com\u002Fkunchenguid\u002Fautopreso.git\ncd autopreso\nnpm install\nnpm start\n```\n\n## How It Works\n\n```\n  ┌──────────┐   audio    ┌──────────────┐   text   ┌──────────────┐\n  │   mic    │──────────► │     STT      │────────► │  whiteboard  │\n  │ (browser)│   24kHz    │ Moonshine \u002F  │ chunks   │    agent     │\n  └──────────┘            │ OpenAI WS    │          │ (OpenAI \u002F    │\n                          └──────────────┘          │  Codex \u002F     │\n                                                    │  Ollama)     │\n                                                    └──────┬───────┘\n                                                           │ tool calls\n                                                           ▼\n                                                  ┌────────────────┐\n                                                  │   Excalidraw   │\n                                                  │  scene (live)  │\n                                                  └────────────────┘\n```\n\n- **Two modes** - \"staging\" lets you sketch seed content client-side; \"live\" hands the canvas over to the agent, biases OpenAI Realtime transcription toward staging text and labels, and starts streaming transcripts.\n- **Local server, local network only** - the Express + WebSocket server binds to 127.0.0.1; nothing is exposed beyond your machine.\n- **Persistent settings** - models, API keys, STT engine choices, and Agent instructions live in `~\u002F.config\u002Fautopreso\u002Fsettings.json` and survive restarts.\n- **Warmup loop** - after you hit start the agent primes itself against your staging content and Agent instructions so the first sentence you say doesn't get a cold model.\n\n## CLI Reference\n\n| Command        | Description                                  |\n| -------------- | -------------------------------------------- |\n| `autopreso`    | Start the local server and open the browser. |\n| `autopreso -h` | Show help.                                   |\n\n### Flags\n\n| Flag         | Description                                   |\n| ------------ | --------------------------------------------- |\n| `--no-open`  | Start the server without opening the browser. |\n| `-h, --help` | Show help.                                    |\n\n## Configuration\n\nSettings persist at `~\u002F.config\u002Fautopreso\u002Fsettings.json` and are managed from the in-app status panel.\nAgent instructions are saved automatically from staging, can be up to 100,000 characters, and take effect on the next Start Preso.\nThe live Session cost card estimates agent token costs and OpenAI Realtime audio costs for the current presentation, resetting on Start Preso or session reset.\nOpenAI prices use the built-in May 2026 rate table; local providers show `$0.0000`, Codex shows token volume because it routes through your subscription, and unknown models show `n\u002Fa`.\n\n### Defaults on first run\n\nWhen no settings file exists, autopreso picks providers based on what it finds in your environment:\n\n| You have...                                | Agent provider                 | Transcription              |\n| ------------------------------------------ | ------------------------------ | -------------------------- |\n| Nothing                                    | OpenAI `gpt-5.5` (needs a key) | Moonshine `medium` (macOS) |\n| `OPENAI_API_KEY` in env                    | OpenAI `gpt-5.5`               | OpenAI Realtime            |\n| Codex CLI signed in (`~\u002F.codex\u002Fauth.json`) | Codex `gpt-5.5-fast`           | Moonshine `medium`         |\n| Codex CLI signed in + `OPENAI_API_KEY`     | Codex `gpt-5.5-fast`           | OpenAI Realtime            |\n| `OLLAMA_MODEL` set                         | Ollama (your model)            | Moonshine `medium`         |\n\nAuto-detection precedence: **Codex CLI auth wins over `OLLAMA_MODEL` wins over `OPENAI_API_KEY`** for the agent. Transcription flips to OpenAI Realtime any time an OpenAI key is present, otherwise Moonshine. After first run, this auto-detection no longer applies - change providers from the in-app status panel.\n\n### Environment variables\n\nProvider variables only seed `settings.json` on first run. Once the file exists, they're ignored - edit the file or use the in-app panel. Log path variables are read on each process start.\n\n| Variable               | Purpose                                               |\n| ---------------------- | ----------------------------------------------------- |\n| `PORT`                 | Port to listen on. Default: `3210`.                   |\n| `OPENAI_API_KEY`       | Seeds the OpenAI key for both agent and Realtime STT. |\n| `OPENAI_MODEL`         | Seeds the OpenAI agent model.                         |\n| `OPENAI_BASE_URL`      | Seeds the OpenAI agent API base URL.                  |\n| `CODEX_MODEL`          | Seeds the Codex model.                                |\n| `OLLAMA_MODEL`         | Seeds the Ollama model.                               |\n| `AUTOPRESO_CACHE_LOG`  | Cache usage log path. Default: `~\u002F.config\u002Fautopreso\u002Flogs\u002Fcache.log`. |\n| `AUTOPRESO_DEBUG_LOG`  | Agent debug log path. Default: `~\u002F.config\u002Fautopreso\u002Flogs\u002Fdebug.log`. |\n\nLocal Moonshine transcription ships as an optional native sidecar for `darwin-arm64` and `darwin-x64`. On other platforms, choose OpenAI Realtime in the STT panel.\n\n## Credits\n\n- [Excalidraw](https:\u002F\u002Fgithub.com\u002Fexcalidraw\u002Fexcalidraw) - the whiteboard canvas, scene model, and rendering.\n- [Moonshine](https:\u002F\u002Fgithub.com\u002Fmoonshine-ai\u002Fmoonshine) the local speech-to-text model that makes the offline path possible.\n- [Vercel AI SDK](https:\u002F\u002Fgithub.com\u002Fvercel\u002Fai) - tool-calling agent loop and provider abstraction.\n\n## Development\n\n```sh\nnpm install                       # install deps\nnpm run dev                       # run the CLI from source\nnpm run typecheck                 # tsc --noEmit\nnpm test                          # node --test\nnpm run build:moonshine-sidecars  # build the Python sidecar binaries\n```\n","autopreso 是一个实时语音转演示文稿的工具，旨在让白板自动绘制内容。其核心功能是通过用户的语音输入，结合自然语言处理模型实现实时的文字转写，并在Excalidraw画布上动态生成图表、标注和布局调整，整个过程无需手动点击操作。该项目采用JavaScript编写，支持用户自定义使用的AI模型（如OpenAI API），并且可以完全本地化运行以保护隐私。适用于需要进行即兴演讲或教学分享等场景，帮助演讲者更专注于内容表达而非幻灯片制作。",2,"2026-06-11 03:32:08","CREATED_QUERY"]