operant
Run tasks in apps with no API: a model discovers the task once on the live screen, then replays it deterministically with human approval on risky...
Tool overview
Overview
Operant is a system for automating tasks in applications that lack APIs. It uses a model to discover how to perform a task by interacting with the screen, then compiles that into a reusable capability that can be replayed without the model. The process consists of three steps: Discover, Compile, and Replay. During discovery, the model drives the real screen through the macOS accessibility tree. The recording becomes a capability with typed inputs and outputs. Replay runs the capability without a model, resuming from the current app state. Security and human oversight are integral. Every risky, mutating, or sensitive step pauses for human approval. Credentials are never exposed to the model; they are requested by name and resolved at replay time. Operant requires Python 3.12 and uv, Node 20 and pnpm for the web UI, and macOS for live runs. A model API key is needed only for discovery; replay needs no key. The system can be tested without live services.
Features
- AI assistant
- Natural language interface
- Workflow automation
Tags
Use Cases
Teams use Operant to automate tasks in applications that have no API.
Teams use Operant to record a task once and replay it without a model.
Teams use Operant to require human approval for risky steps.
Teams use Operant to handle credentials securely without exposing them to the model.
Teams use Operant to resume tasks from the current app state.
Teams use Operant to test automation without live services.
User Reviews
Related Tools
View moreLynote
Lynote combines AI-draft rewriting and AI-text detection with note-taking, transcription, video summaries, and study tools.
LMSYS Chatbot Arena Leaderboard
LMSYS Chatbot Arena is a crowdsourced open platform for LLM evals. Collected over 1,000,000 human pairwise comparisons to rank LLMs with the Bradley-Terry model and display the model ratings in Elo-scale.
智谱清言
Zhipu Qingyan is a GLM-based AI assistant whose official description emphasizes understanding goals, breaking down tasks, and using tools.
豆包
字节跳动AI助手