taskuary

# Taskuary [English](README.md) · [简体中文](README.zh-CN.md) [](https://github.com/ldbumble/taskuary/actions/workflows/ci.yml) [](https://scorecard.dev/viewer/?uri=github.com/ldbumble/taskuary) [](https://pypi.org/project/taskuary/) [](https://github.com/ldbumble/taskuary) [](LICENSE) [](https://github.com/ldbumble/taskuary/stargazers) <p align="center"><b>⭐ Please star Taskuary if you find it helpful</b> — it is how other people find it.</p> ## Your inbox, staffed by AI agents Taskuary turns incoming messages into organized work. It sorts what matters, hands tasks to your agents, and brings decisions back to you. Nothing sends or ships without your approval.  Taskuary is early—currently **v0.3.7.6**—so breaking changes are still possible before 1.0. <p align="center"> <a href="https://taskuary.com/demo/"><img src="https://img.shields.io/badge/%E2%96%B6%20Try%20it%20now-no%20install%2C%20in%20your%20browser-2f4858?style=for-the-badge&labelColor=1f2a22" alt="Try Taskuary now, in your browser"></a> </p> <p align="center">_{The real app with invented data. Nothing connects, sends, or runs.}</p> <p align="center"><b><a href="https://taskuary.com/docs/">Read the documentation</a></b> — installation, the first run, what to connect, and every setting.</p> ## What Taskuary can do One request, from arrival to your approval. Follow Ruth's request for the latest vendor spend numbers through the real app, using fictional demo data. ### 1. Everything lands in one place Mail, chats, issue trackers, alerts, and reports arrive on one rail, sorted by what they need: **Urgent**, **On you**, **For later**, the Advisor's ideas, and **FYI**. Beside it, the Assistant says what came in and who is waiting, so you never open each system in turn.  ### 2. Let the Assistant walk you through it Choose **Walk me through my tasks**. The Assistant puts one item on the table at a time, as the same task card you would open yourself, with the next actions under it: **Next**, **Write reply**, **Start an agent**. The buttons always do the same thing; the chat line is for everything else, and the Assistant already has the message, the thread, and your past decisions to answer from.  ### 3. Every request becomes a task with its whole story Open any item to see how it got here: the message, what triage decided and why, which agent has it, and where the reply stands.  ### 4. The agent works inside the chat **Start an agent** and it works in the same card. Here the analyst prepares the numbers, checks that the categories add up, names its source, and drafts the reply. Nothing is sent.  ### 5. The last word is yours The prepared reply waits on the task, beside the request that started it. Read it, edit it, and choose **Approve & send** when it is ready.  ### 6. Start your day before it gets busy Each morning the Assistant opens with today's meetings on a timeline and who needs a word from you. Ruth's request has a clear deadline: the 11:30 operations review.  ## Key features ### Use your coding CLI Connect Claude Code, Codex, Qwen Code, OpenCode, Kimi Code, Gemini, Cursor, Copilot, Muse Code, or another CLI. Set up the connection once, then give your agents profiles with their own instructions. Follow their sessions, answer questions, and review the result from Taskuary. For Qwen Code, see the [setup and compatibility guide](docs/qwen-code.md). Use DeepSeek, GLM, or MiniMax through OpenCode, or connect Moonshot's Kimi Code: [setup steps and supported roles](docs/chinese-coding-clis.md).  ### A shared Hub for what agents learn Keep discoveries, decisions, and useful warnings by topic. Agents can find what earlier work uncovered, discuss it, and correct it instead of starting from scratch.  ### Agents leave notes for each other The Board's **Live handoffs** show what agents are working on, what is blocked, and what is ready. An agent leaves a note; the next one reads it before picking up the work.  ### Memory that learns how you work Every draft you edit and task you reclassify gives Taskuary evidence about your preferences. Repeated patterns become lessons in **LEARNED.md**: how you write, what you own, and what deserves a task. For example, repeatedly moving the numbers to the top of a reply can teach it to lead with the total next time. Open **Docs → LEARNED.md** to read, edit, or delete those lessons. Your written instructions in `SOUL.md` take precedence.  <details> <summary>Technical details: how LEARNED.md becomes memory</summary> - **Learn from a correction.** A model call turns an explicit correction into a hypothesis. Batched reflection compares multiple decisions; untouched approvals contribute here. - **Keep the evidence.** Each machine-written lesson carries `s` (strength), `ev` (evidence IDs), and `seen` (last supporting date). A stable `k` identifies it across rewrites. - **Promote supported patterns.** A new hypothesis starts at strength 2. Reflection is instructed to promote it at 4 or more, with at least three episodes across two people or threads. Contradictions weaken it; stale hypotheses also decay across reflection cycles. - **Use active lessons.** The prompt builder excludes the Hypotheses, Proposed rules, and raw Verdicts sections. Active lessons inform triage, drafts, and agent context. - **Keep the owner in charge.** Inferred rules that hide or file work wait in Proposed rules. Two matching explicit owner verdicts can already supply that authorization. Untagged lines you write are preserved, and learning can be disabled in Settings. See [Learning from your decisions](https://taskuary.com/docs/how-it-works#correcting-it-teaches-it), or the implementation in [learn.py](taskuary/learn.py) and [learnedgraph.py](taskuary/learnedgraph.py). </details> ### What leaves your machine  Taskuary runs locally. With a hosted AI provider or coding CLI, the prompt contains the context needed to do the work you asked for. **Use a model running on your machine, and those AI prompts stay local too.** Credentials are taken out of that prompt first. If a colleague mails an API key, a connection string, or a private key, it is replaced with a labelled placeholder (`[redacted:aws-key]`) at each of the three doors a prompt can leave by: the hosted models, a headless CLI run, and the first prompt of an agent pane. Your mail itself is never altered—the scrub is on the way out, not on the way in—so a vendor's one-time code stays readable where it arrived. Nothing Taskuary sends carries a placeholder either: a reply still holding one is refused, not delivered. The rules are deterministic rather than a model's judgement, because by the time a model could judge, the credential would already be in a prompt. So they catch credentials with a recognizable shape—provider keys, tokens, credentialed URLs, connection strings, private keys—and they will not catch a sentence like "the wifi password is bluefish17". Report anything you find through [SECURITY.md](SECURITY.md). ## Install ### Windows app Download the latest single-file [Taskuary.exe](https://github.com/ldbumble/taskuary/releases/latest/download/Taskuary.exe) and open it. No Python or installer is required. ### Python Python 3.10 or newer works on Windows, macOS, and Linux: ```bash pip install taskuary taskuary ``` Taskuary opens at [http://127.0.0.1:7787](http://127.0.0.1:7787). For a native desktop window instead, install `pip install "taskuary[desktop]"` and run `taskuary-desktop`. ### Docker ```bash git clone https://github.com/ldbumble/taskuary cd taskuary docker compose up ``` Then open [http://127.0.0.1:7787](http://127.0.0.1:7787). Docker runs the web app; coding CLIs and the optional WhatsApp bridge remain on the host. On first run, connect an AI provider or local Ollama model, add at least one inbound channel, then choose the coding CLI that should receive tasks. The setup wizards test each connection before it goes live. ## Try it without installing anything ```bash taskuary --demo # or: docker compose --profile demo up ``` The demo is the real interface with fictional work and scripted replies. It cannot connect to outside systems, send messages, run tools, or start agents. Its changes reset when you reload. ## Installs  Updated daily from PyPI with mirror traffic excluded. The raw series is [downloads.csv on the `stats` branch](https://github.com/ldbumble/taskuary/blob/stats/downloads.csv). ## Documentation The documentation is at **[taskuary.com/docs](https://taskuary.com/docs/)**. - [Start here](https://taskuary.com/docs/)—installation, first run, Docker, and where your data lives - [How it works](https://taskuary.com/docs/how-it-works)—the Timeline, the five roads, what triage decides, the operator documents - [Connections](https://taskuary.com/docs/connections)—channels, AI providers, work systems, and report sources - [Tasks and agents](https://taskuary.com/docs/tasks-and-agents)—the task, the agent work and the reply as three separate lives - [Reports and the Assistant](https://taskuary.com/docs/reports)—the report pipeline, AI-written source cards, and what Taskuary watches - [Settings reference](https://taskuary.com/docs/settings)—every setting, generated from the schema the app itself reads - [Status and roadmap](docs/roadmap.md)—what works today and what is next - [Contributing](CONTRIBUTING.md)—development setup and contribution guide Taskuary is free and open source under the [MIT License](LICENSE). Issues and pull requests are welcome; security reports belong in [SECURITY.md](SECURITY.md).