dengkui yang

We started building this after seeing how much energy gets lost when an important workflow depends on too many disconnected tools and too much manual cleanup. With MarkItDown, teams can turn documents, webpages, tables, and structured data into clean Markdown for review, search, and LLMs. We wanted that kind of work to feel more practical in everyday use, easier to explain to a team, and easier...

We started building this after seeing how much energy gets lost when an important workflow depends on too many disconnected tools and too much manual cleanup. Primary focuses on onyx turns docs, tickets, code, chat, and policy pages into one permission-aware knowledge layer with citations, freshness, and flexible deployment behind every answer. We wanted that kind of work to feel more practical...

We started building this after seeing how often teams had to trust a filename, an extension, or a quick guess when deciding what an uploaded file really was. Magika reads the file content and returns a detected type, MIME, file group, confidence score, and extension mismatch signal. We wanted that kind of work to feel more practical in everyday use, easier to explain to a team, and easier to...

We started building this after seeing how much energy gets lost when an important workflow depends on too many disconnected tools and too much manual cleanup. With ML Intern, we wanted to create a turn an ML task into a usable run brief shaped around papers, Hub search, GitHub inspection, and HF Jobs that feels more practical in everyday use, easier to explain to a team, and easier to learn...

We started building this after seeing how much energy gets lost when an important workflow depends on too many disconnected tools and too much manual cleanup. With OpenClaude, we wanted to create an openClaude gives teams one coding-agent workflow across OpenAI-compatible APIs, Gemini, GitHub Models, Codex, Ollama, MCP, and VS Code that feels more practical in everyday use, easier to explain to...

We started building this after seeing how much energy gets lost when an important workflow depends on too many disconnected tools and too much manual cleanup. With CC Switch Cloud, we wanted to create a switch providers, MCP servers, prompts, skills, and sessions across Claude Code, Codex, Gemini CLI, OpenCode, and OpenClaw that feels more practical in everyday use, easier to explain to a team,...

Hey Product Hunt! I built Claw Simple to make launching OpenClaw on Telegram much less messy. A lot of the friction was never just the bot itself — it was the setup around it: hosting, runtime choices, billing, and getting everything launch-ready without jumping between tools or living in the terminal. Claw Simple is my attempt to turn that into one clearer flow. Pick a plan, choose an AI mode,...

Hey Product Hunt — I built CoPaw for people who want more than a chat window. A lot of AI tools feel useful at first, but once you need memory, repeatable workflows, channels, rules, or actual operating context, things get messy fast. I wanted a calmer setup that could bring those pieces together in one place. CoPaw is my attempt at that: an open-source personal AI workstation with local setup,...