Midscene
An open-source vision-based UI automation toolkit for writing natural-language tests and actions across web, desktop and mobile interfaces.
About Midscene
Developers need to test and automate interfaces where DOM selectors or accessibility metadata do not cover the required interactions.
Who it’s for
- Engineers writing end-to-end UI tests
- Developers automating web and native applications
When to consider it
Consider Midscene when you want screenshot-based UI actions, natural-language assertions and visual execution reports in your automation tests.
Tradeoffs & limitations
- Midscene is MIT licensed; running automation requires configuring a supported multimodal model endpoint, model name, API key and model family.
SOTA overview · Documentation-based assessment · Sources & review method
Updates
No updates shared yet.
Discussion
Newest firstAsk a question or share how you use Midscene.
Keep it helpful. Community rules
Loading discussion…
Sources & review method
Documentation-based assessment · Oct 5, 2026 · Prepared with AI assistance; not a hands-on benchmark.
- Midscene official website
- Midscene model configuration
- Midscene repository
- Midscene MIT license
- Midscene September 29 commit
Checked by SOTA · AI-assisted documentation review. Selection advice is our assessment; verify current requirements for your deployment.
Import history & original evidence
Midscene official website
An open-source vision-based UI automation toolkit for writing natural-language tests and actions across web, desktop and mobile interfaces.
Based on official pages and announcements checked on 2026-10-05. No hands-on test, performance benchmark or popularity ranking is claimed.