AI Tool Podium

Issue #1 ยท

Kimi K3.1 Pro, Cursor computer use, and 14 new tools ranked

Once a week: what launched, what we added to the rankings, and the current top pick for each job.

This is the web edition of the AI Tool Podium email, and the first issue. Ranks are as of Oct 24, 2026; each tool page shows its current position.

Worth knowing this week

  1. Moonshot AI released Kimi K3.1 Pro, which it calls its most capable 3T-class open-weight model, with a 1-million-token context window and gains in coding, 3D creation and agent tasks.

    Source
  2. Cursor agents on self-hosted machines can now click, type, take screenshots and drive desktop apps and browsers on macOS and Linux workers, with optional desktop sharing on Linux.

    Source
  3. Anthropic's research preview of a proactive agent app: a daily brief from your connected apps, background tasks on its own virtual machine, and approval before it sends, pays, signs or deletes.

    Source
  4. Resend announced Inboxes: shared email inboxes on your own domain, built for agents and teams, with every message grouped into threads your app and agents can read and answer.

    Source
  5. Mistral's new speech family in public preview: Mini Transcribe 3, Mini Transcribe Realtime 3, Small Transcribe 3 and TTS 3, plus a Voice Agents SDK built on LiveKit.

    Source
  6. Exa's biggest search-quality update, with a larger embedding model, Dynamic Highlights now generally available, and an expanded index.

    Source

New on the podium

14 added
  1. CrawlForge#10 in AI agents and automation

    Open-source MCP server with 31 web tools (scrape, crawl, stealth browse, multi-source research) for Claude Code, Cursor and other agents.

    Read
  2. MemHub#11 in AI agents and automation

    Tencent Hunyuan's local-first memory app for agents.

    Read
  3. Praex#8 in Terminal coding agents

    Terminal coding agent built on a fork of OpenCode (MIT) that runs on Praex's own hosted models: Velox II free every day, Faber II on the $20/month Pro plan, or bring your own keys.

    Read
  4. Composio#12 in AI agents and automation

    Tool and auth layer for agents: gives Claude, Codex, Cursor and other MCP clients access to 1,500+ apps through SDKs, a CLI and an MCP gateway.

    Read
  5. TwinMind#5 in AI meeting notes

    Captures and summarizes meetings, calls and conversations on iOS, Android, Mac, Windows and Apple Watch, then lets you chat with that context.

    Read
  6. Harness by Autonomous#13 in AI agents and automation

    Free, open-source (MIT) command center that runs Claude Code, Codex, Cursor, OpenCode and other coding agents side by side across your machines, each on its own git worktree, with end-to-end encrypted links.

    Read
  7. Croo#9 in Terminal coding agents

    Tiny MIT-licensed terminal agent with zero runtime dependencies.

    Read
  8. heal.dev#5 in AI code review

    Runs your app in a production-like sandbox and files only bugs it can reproduce, each with a failing test.

    Read
  9. Vidmoat#5 in AI video generation

    Browser video editor where an agent cuts, captions and grades your footage on a real timeline you can still edit by hand.

    Read
  10. fx#10 in Terminal coding agents

    Vercel Labs' open-source (Apache-2.0) terminal agent shipped as one small native binary for macOS and Linux.

    Read
  11. Mistral Voxtral 3#5 in AI voice and audio

    Mistral's speech models for developers: batch and realtime transcription, text-to-speech and a voice agents SDK.

    Read
  12. PhoneUse#14 in AI agents and automation

    Gives AI agents a real phone to drive mobile apps through MCP or an API: cloud Android phones billed monthly or per minute, plus a robot-tapped iPhone sold by call.

    Read
  13. Undermind#5 in AI research and notes

    AI literature search agent that runs deep searches across papers and writes reports; free plan, Pro from $16 a month billed annually.

    Read
  14. Semantic Scholar#6 in AI research and notes

    Free AI-powered academic search from AI2 that indexes over 200 million papers.

    Read

Top picks by use case

  1. CursorVibe coding (AI code editors)

    AI-first code editor with agent mode; the default pick for building apps by prompting.

    Read
  2. Claude CodeTerminal coding agents

    Terminal agent that is strong at multi-step changes across a codebase.

    Read
  3. LovableNo-code app builders

    Prompt-to-full-stack-app builder; easiest path from idea to working web app.

    Read
  4. ChatGPTGeneral AI assistants

    The most versatile all-round assistant, with search, files, voice and images.

    Read
  5. RunwayAI video generation

    Pro-oriented video generation and editing tools for short clips.

    Read
  6. MidjourneyAI image generation

    The most stylish, artistic image output.

    Read

Head to head

  1. Read
  2. Read
  3. Read

Some placements on AI Tool Podium are sponsored. Every sponsored placement is clearly labeled. All issues

The weekly podium

Ranking changes and new releases, once a week.

The tools that moved, the launches worth knowing about, and new head-to-head comparisons. Read issue #1.

  • Free
  • One email a week
  • Unsubscribe anytime