Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Tactile: an accessibility-first operating layer for agents.
| Date | Stars |
|---|---|
| 2026-07-31 | 614 |
| 2026-08-06 | 614 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
<p align="center"> <img src="assets/logo.png" alt="Tactile logo" width="160"> </p> <p align="center"> <strong>English</strong> · <a href="README_zh.md">简体中文</a> </p> # Tactile **An accessibility-first operating layer for agents.** > Stop guessing pixels. Start touching semantics. Tactile is not another computer-use agent. It is a skill, protocol, and tool layer that helps agents operate software through accessibility semantics first. When an agent needs to use an application, Tactile asks it not to begin with screenshots, guessed coordinates, and pixel-level clicks. Instead, it should first inspect the semantic information already exposed by the operating system and the application: - What role does this element have? - Does it have an accessible name? - Is it clickable, selected, focused, enabled, or disabled? - Where does it sit in the UI hierarchy? - Does it expose an action that can be invoked directly? In that sense, Tactile gives agents a way to feel the structure of software before reaching for vision. This information already exists for screen readers and assistive technologies. Tactile makes it the first entry point for agents as well. The easier software is for Tactile to operate, the more likely it is to support genuinely accessible interaction for humans as well. **Agent-ready software should also be human-accessible software.** ## Demo **Tactile gives agents a sense of touch.** ### Lark and WeChat workflow This demo video was also edited by an agent using the Tactile skill to operate CapCut. https://github.com/user-attachments/assets/49dc6bfe-0661-4ab0-9099-be3849b4137a ### CapCut video-editing workflow https://github.com/user-attachments/assets/7bc0f05e-9228-4cf1-abe3-ffb7e4722be2 ## How to Use Tactile ### Prefer the macOS MCP For macOS, you can use the dedicated Tactile MCP in `mcps/tactile-macos-mcp`. It is the recommended entry point when available: it is faster, easier to use, and exposes Tactile's accessibility-first workflow directly through MCP tools. Use the MCP server with: ```bash mcps/tactile-macos-mcp/bin/tactile-macos-mcp ``` You can also directly ask your agent (Codex / Claude Code) to install this MCP: https://github.com/yliust/Tactile/tree/main/mcps/tactile-macos-mcp ### Use Tactile as a skill Ask your agent to configure this skill from the repository: ```txt Configure this skill for me (make sure to choose the version for the corresponding operating system): https://github.com/yliust/Tactile ``` If using API: ```txt export TACTILE_OPENAI_BASE_URL=xxxxxxx export TACTILE_OPENAI_API_KEY=xxxxxxx export TACTILE_MODEL='gpt-5.5' ``` ## Why Tactile? Many computer-use agents start from screenshots: ```txt look at screenshot -> infer element -> predict coordinates -> click -> inspect screenshot again ``` This approach is general, but fragile. Tactile changes the order of operations: ```txt read accessibility semantics -> use OCR-grounded coordinates when needed -> fall back to visual computer use ``` Agents should not only see software on a screen. When better information is available, they should first touch the structure of the interface. ## Tactile v0 Tactile v0 will begin as a skill. Its goal is to package an accessibility-first operating method for agents: 1. **Use accessibility semantics first** If the system or application exposes useful accessibility information, the agent should use element roles, names, hierarchy, state, and actions to understand and operate the interface. 2. **Use OCR + coordinates when semantics are incomplete** If an element is not fully represented in the accessibility layer but the visible text is readable, the agent can use system OCR. System OCR usually returns both text and coordinates, which makes it a text-grounded fallback rather than pure visual guessing. For clear text buttons and labels, this can reduce token usage, retries, and time. 3. **Fall back to the agent's native visual operating logic** If the acce
Excerpt of 9,094 characters
Read on GitHub
30
7
3
2
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:15842aca0722b5e0, llm:Repository description: "Tactile: an accessibility-first operating layer for agents." (no topics, Python)
matched fp:15842aca0722b5e0, llm:Repository description: "Tactile: an accessibility-first operating layer for agents." (no topics, Python)
matched fp:15842aca0722b5e0, llm:Repository description: "Tactile: an accessibility-first operating layer for agents." (no topics, Python)