Skip to content

Commit 907ac6d

Browse files
authored
Update README.md
1 parent 5fd5d91 commit 907ac6d

File tree

1 file changed

+23
-9
lines changed

1 file changed

+23
-9
lines changed

README.md

Lines changed: 23 additions & 9 deletions
Original file line numberDiff line numberDiff line change
@@ -32,22 +32,36 @@ https://github.com/user-attachments/assets/00329105-8875-48cb-8970-a62a85a9ebd0
3232
</a>
3333
</p>
3434

35-
⚠️ News:
36-
- 08/25 shipped tons of new features & bug fixes: run NodeJS SDK in YAML workflows, run JS in browser, record OS events and genereate YAML workflows directly in MCP end-to-end, etc.
35+
## 🚀 What's new
36+
37+
- 08/25 Big release — NodeJS SDK in YAML workflows, run JS in browser, OS event recording → YAML generation in MCP, and more
3738
- 08/25 [we raised $2.8m to give AI hands to every desktop](https://x.com/louis030195/status/1948745185178914929)
3839

40+
> AI-powered desktop automation — open source, MIT-licensed, cross-platform, accessibility-first.
41+
Works across all apps and browsers. Inspired by GitHub Actions & Playwright.
42+
100x faster than generic AI agents, >95% success rate, no vendor lock-in.
43+
44+
## 🧠 Why Terminator
45+
46+
### For Developers
47+
- Create automations that work across any desktop app or browser
48+
- Runs 100x faster than ChatGPT Agents, Claude, Perplexity Comet, BrowserBase, BrowserUse
49+
- \>95% success rate without costly consultant armies (UIPath, Automation Anywhere)
50+
- MIT-licensed — fork it, ship it, no lock-in
3951

40-
> Computer use SDK for building agents that learn from human screen recordings. Accessibility-first. Cross-platform (Windows/macOS/Linux).
52+
We achieve this by pre-training workflows as deterministic code, and calling AI only when recovery is needed.
4153

42-
There are three paths to train deterministic workflows with AI fallback:
54+
### For Teams
55+
[Our private beta workflow builder](https://www.mediar.ai/) + managed hosting = plug-and-play automation engine for your business:
56+
- Record & map your processes in days
57+
- Deploy AI to execute them at >95% accuracy
58+
- Kill repetitive work without legacy RPA complexity or cost
4359

44-
1. **AI-Assisted Workflow Building**: Use an MCP client like [Cursor](https://cursor.com) to iteratively build and test complex workflows with an AI assistant in a human-in-the-loop process.
45-
2. **Record Human Baselines**: Use our open-source tools to record a human demonstrating a task (our MCP has a recording tool). This generates a baseline workflow that can be refined and automated.
46-
3. **Enterprise-Grade Recording**: For businesses needing scalable, >95% success rate automation workflows, our [b2b product](https://mediar.ai) provides the most robust solution.
60+
## ⚡ Quick start: MCP
4761

48-
For detailed instructions on building with AI agents through MCP client, see our [**Terminator MCP Agent README**](terminator-mcp-agent/README.md).
62+
We recommend using [MCP](https://github.com/mediar-ai/terminator/tree/main/terminator-mcp-agent) as AI's IQ is far higher than the programmer.
4963

50-
## ⚡ Quick Start: Programmatic Control
64+
## ⚡ Quick start: programmatic control
5165

5266
### 🐍 Python
5367

0 commit comments

Comments
 (0)