askimo-ai/askimo

What it solves

Askimo is a native desktop application that eliminates the need to switch between multiple browser tabs and AI providers. It solves the problem of fragmented AI workflows by consolidating various LLMs, local file search (RAG), web search, and autonomous agent execution into a single, privacy-focused interface where data remains on the user's machine.

How it works

Askimo acts as a unified hub that connects to multiple AI providers (such as OpenAI, Claude, Gemini, and local options like Ollama or LM Studio) via API keys or local endpoints. It implements a local RAG system using hybrid BM25 and vector retrieval to index local files and URLs without uploading them. For complex tasks, it uses a "Plans" engine to chain multi-step AI pipelines and integrates with Model Context Protocol (MCP) servers and autonomous CLI agents (like Claude Code) to execute real-world tasks.

Who it’s for

It is designed for power users who want a high-performance desktop experience for AI, those who prioritize data privacy and local storage, and developers who need to orchestrate multi-step AI workflows and autonomous agents without being locked into a single provider.

Highlights

• Multi-Provider Support: Seamlessly switch between cloud LLMs and local models (Ollama, vLLM, etc.) within a single session.
• Local-First Privacy: Conversation history, RAG indices, and telemetry are stored locally in SQLite; no data is uploaded to the project maintainers.
• Agentic Workflows: Create multi-step "Plans" via a form UI or YAML, and run autonomous coding agents with live tool-call streaming.
• Extensible Tooling: Integration with MCP servers and a built-in script runner for Python, Bash, and JavaScript.
• Local Voice Capabilities: Supports offline speech-to-text (Whisper-compatible) and text-to-speech (Piper).