Common
Nous Research's open-source agent: runs on your own machine, keeps long-term memory, turns past work into reusable skills and answers you in Telegram, Slack and more. Bring your own model.
Tap the card to see all five ratings
Common · Does one simple job well enough
Hermes Agent
Nous Research's open-source agent: runs on your own machine, keeps long-term memory, turns past work into reusable skills and answers you in Telegram, Slack and more. Bring your own model.
- Score
- 80
- Rarity
- Common
- In China
- Runs locally
- Payment
- Free & open source
- Category
- Agents
Five ratings
Score = output quality 30% + efficiency 20% + ease of use 15% + reliability 20% + cost 15%. Higher is better on every rating, compared within the same category. A higher cost rating means cheaper.
Updated 2026-10-31
Similar tools
Tencent’s desktop office agent. Describe the job and it tidies files, cleans spreadsheets and builds reports and slides, and you can drive it remotely from WeCom or QQ.
Meta’s personal AI agent. Connect your email, calendar and shopping accounts and it books restaurants and trips and places orders. US only for now.
Moonshot AI’s local work agent. Give it a goal and it splits the job, runs sub-agents in parallel for research, organizes files and writes reports. In public beta.
ByteDance’s agent builder with drag-and-drop workflows, publishable to Douyin and Lark.