Social content creation, made simple
A desktop content studio with four workbenches — AI ideation, material organization, WeChat article formatting, and social card generation. Typesetting runs on a local Rust engine, not a model.
What it does
Each capability maps to a concrete everyday problem.
Ideation: AI chat
Three capability groups — brainstorming, writing, and organizing — with built-in prompts for articles, headlines, polishing, and topic planning. One click fills the prompt box.
- Combine skills, knowledge base, and model per chat
- Conversation history panel for quick switching
Materials: WeRead import
Import WeRead highlights and notes in one click, filter by book or entry, and export as Markdown straight into the article or card workbench.
- API keys live in the OS keychain, not plain files
- Notion / Obsidian entry points planned
Articles: WeChat formatting
Markdown on the left, live preview on the right. Switch between 22 presets instantly, with title tags, author info, table of contents, and keyword highlighting.
- Copy to WeChat or export HTML in one click
- Rule-based or AI keyword extraction
Cards: Social media cards
Split source material into 1–10 cards, manually or with AI generating the whole set at once. Switch between carousel and grid views to check the overall look.
- AI style transfer: sketch, oil, watercolor
- Export one card or the whole set as PNG
Bring your own model
Works with Ollama, OpenAI, Anthropic, DeepSeek, GitHub Copilot, MiniMax, SiliconFlow, StepFun, OpenRouter, NVIDIA, Bailian, Volcengine, and custom endpoints.
- Fallback chains switch in order when a model fails
- Failed nodes cool down with exponential backoff
Deterministic engine
Layout, table of contents, numbering, and validation run locally in Rust — stable, reproducible, and free of token cost.
How it works in practice
Not a feature list — a walk through getting something done.
From one sentence to a clear draft
Three capability groups with built-in prompts that fill the input box in one click, so a rough idea becomes a first draft fast.
Combine a skill, a knowledge base, and a model next to the input box to match your writing habits. The history panel keeps every conversation within reach.
Write Markdown, get publish-ready layout
The toolbar handles drafts, Markdown import and export, image insertion, and the usual inline formatting.
Switch theme tabs to apply one of 22 presets across font sizes, colors, and emphasis. The author tab sets byline and avatar; the misc tab controls TOC, numbering, and keyword highlighting.
One source, a full set of cards
Paste material or import from WeRead, and AI splits it into titles, subtitles, and body text for 1–10 cards. Switch back to manual mode anytime to fine-tune each card.
Upload an image on any card and convert it to pencil sketch, oil painting, or watercolor, with content presets and customizable canvas size and palette.
Run locally or plug in your provider
Fill in a base URL and API key for any of the supported providers, fetch the model list, and test the connection before you start writing.
The fallback tab lets you chain backup models. When the primary fails, the next one takes over, and the failed node cools down with exponential backoff.
Design trade-offs
Deterministic work stays in code; judgement stays with you.
Native Rust layout engine
Formatting, TOC generation, and section numbering run locally — no model involved, fully reproducible.
Swap providers freely
Local Ollama or 12+ cloud providers, with fallback chains so one outage does not stop you.
Credentials stored safely
API keys for WeRead and model providers live in the OS keychain, not in plain text files.
Native Tauri desktop
Built on Tauri, so native file dialogs and local AI proxies work as expected.
Local knowledge retrieval
Import Markdown or JSON to build a vector knowledge base that grounds answers in your own material.
Cross-session memory
Manage memory across sessions so you do not have to re-establish context every time.
What you might want to know first
Where are my API keys stored?
Provider and WeRead keys are stored in local app storage or the OS keychain. That suits personal machines; avoid keeping sensitive credentials on shared accounts.
Does it need an internet connection?
No. Layout, numbering, and validation run locally in Rust. Pair it with a local Ollama model and the creative features work fully offline too.
Which platforms are supported?
A native desktop app built with Tauri, available for Windows and macOS.
Is it free?
Downloading and using it is free; tips are optional. AI features require your own model API. The source is not open yet — only release installers are provided.
Start using it now
Free to download, runs locally. Your data stays on your machine.
This is a desktop app — open this page on your computer to download.
Other formats
SHA256 checksum
Provided as-is. The app does not collect usage data.