cllama Autor: cllama

cllama puts an AI sidebar in your browser: understand the page you are on, ask it anything, and get real work done — without switching tabs.

📱 cllama for Android
💻 github.com/trytop10/cllama

• One-click page insight — use a built-in recipe, or right-click a selected passage → Insight, to turn any page into a summary, key points, term explanations or an argument analysis. Results are saved as Markdown cards you can fold, copy and revisit; write your own recipes and pin the exact content area per site.

• One client for all your models — use it as a local Ollama client (free, private, data stays on your machine), or connect OpenAI, Gemini and any OpenAI-compatible remote service. Keep several configurations, switch in a click, or type @ to use another model for one message.

• Skills that actually do things — page Q&A, date & time, or workflows you write yourself. Page Builder turns the current page (or the current conversation) into a reusable Skill, chat scenario or insight recipe — imported only after you confirm.

• MCP, including MCP Apps — connect MCP servers to use their remote tools, and when a server ships an MCP App its interactive UI is rendered right inside the chat, so you can click instead of typing.

• Generate and edit images — text-to-image and image-to-image with supported vendor models (e.g. OpenRouter image models): pick generated results as references for the next turn and keep working alongside images you upload.

• Nothing happens behind your back, and the chat stays tidy — writing data or calling an external service asks for confirmation and leaves a readable run log; scenarios, per-conversation skill memory, search in one or all conversations, saved answers, HTML/PDF export, token usage per reply and attachments are all built in.

• Gets better as you go — press Reflect in a scenario and it distils the conversation into reusable notes; you pick what to keep, and later turns use them automatically.

Free, no account needed, 15 UI languages.