Compact Rust autonomous assistant runtime optimized for speed and minimal resource usage.
Shrew is a compact autonomous assistant runtime written in Rust, optimized for speed and minimal resource usage. It is an experiment in bringing agent loops to systems languages for fast, low-overhead personal assistants. Very early stage, but an interesting systems-level take on agent runtimes.
| Pricing | Open Source |
|---|---|
| Category | Productivity |
| Platforms | linux, macos |
| Official website | github.com |
When choosing a productivity tool, look first at how it fits the way you already work. The best tools connect with the apps you use every day — email, calendar, documents and chat — instead of becoming one more place to check. Automation matters most where work repeats: summaries, reminders, scheduling and turning notes into tasks are where AI genuinely saves time. Check whether the free tier covers your real usage, how the tool handles your data and privacy, and whether it works on the platforms you live on — web, desktop and mobile. A slightly less powerful tool that your whole team actually adopts beats a feature-rich one nobody opens.
Shrew is a compact autonomous assistant runtime written in Rust, optimized for speed and minimal resource usage. It is an experiment in bringing agent loops to systems languages for fast, low-overhead personal assistants. Very early stage, but an interesting systems-level take on agent runtimes.
Shrew offers Autonomous assistant runtime in Rust, Minimal memory and CPU footprint, Fast startup and low latency, Tool-calling agent loop and Small, auditable codebase.
Shrew is open source, so the software itself is free to download, use and self-host. Your only costs are the infrastructure you run it on and any paid services you connect to it.
Strengths of Shrew include Extremely lightweight by design; Rust performance and safety; Novel systems-language agent runtime. On the downside, Tiny project (9 stars) with little documentation; Rust + LLM integration is still experimental; No integrations or community ecosystem yet.
by Linkly AI
Browser extension from Linkly AI that saves web pages as agent-friendly Markdown into your local, private knowledge library, so connected AI agents can search, compare and cite your saved material.
Best for: People building a private knowledge library that their AI agents can search and cite
Pricing: Freemium
Limitation: Paid plans needed for frontier AI models, cloud access and higher quotas
by Modelpeers
Follow AI models and agents like you follow people — it watches changelogs, deprecation notices, status pages and pricing pages, and sends one Monday email when something you rely on changes.
Best for: Anyone whose work breaks when a model, agent or price they rely on quietly changes
Pricing: Free
Limitation: Coverage is 163 models and 31 agents — a tool outside that list is not watched
by magpie
One visual library for everything saved across social platforms, with AI tagging — search by what's in a save or by colour, and keep a real copy even when the original post is deleted.
Best for: People who save constantly across social apps and want to actually find things again
Pricing: Freemium
Limitation: Free tier limited to 500 saves
by BaseTab
Browser new-tab dashboard with 30+ widgets and an AI widget builder, plus Claude, ChatGPT or Gemini in a widget.
Best for: People who want their browser's new tab to be a working dashboard, not a blank page
Pricing: Freemium
Limitation: Free plan limited to one workspace
by Godwit
Open-source AI-native desktop email client for IMAP/SMTP — assistant answers questions about your mail, reads attachments, summarises, translates and drafts with your chosen model.
Best for: People who want an AI-native email client that works with any mail account and any model
Pricing: Open Source
Limitation: You supply your own model or use local models — no model included
by Scape
Intelligent desktop email client that surfaces what matters and prepares replies from your email, meetings, calendar and attachments — you review and send.
Best for: Professionals who live in their inbox and want replies drafted from everything around the email
Limitation: Preview release, with no pricing shown on the official site
