An open-source framework for intelligent browser agents.
Give your agent a goal. Let it navigate the web.
Aevora is built around a simple question:
What if interacting with the web started with intent, not selectors?
Traditional automation asks you to describe every step. Aevora is designed to let an AI agent understand a goal, observe a browser, and execute the actions needed to accomplish it.
ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
β β
β "Research the best options and compare them." β
β β
ββββββββββββββββββββββββββββ¬ββββββββββββββββββββββββββββββββ
β
βββββββββββββββββββ
β AEVORA CORE β
β β
β Observe β
β Reason β
β Act β
ββββββββββ¬βββββββββ
β
βββββββββββββββββββ
β Browser Layer β
ββββββββββ¬βββββββββ
β
βββββββββββββββββββ
β The Web β
βββββββββββββββββββ
Aevora is designed for goal-oriented browser automation, where the agent can interact with websites instead of relying entirely on rigid, predefined workflows.
Give the agent a task and let it work through the interface.
"Fill in this application using my information."
Turn web pages into useful, structured information.
"Collect the relevant data and organize the results."
Move beyond individual clicks and into multi-step browser tasks.
"Research, compare, and return the best options."
Aevora follows an agentic execution loop:
ββββββββββββββββ
β User Goal β
ββββββββ¬ββββββββ
β
βΌ
ββββββββββββββββ
β Understand β
β Intent β
ββββββββ¬ββββββββ
β
βΌ
ββββββββββββββββ
β Observe β
β Browser β
ββββββββ¬ββββββββ
β
βΌ
ββββββββββββββββ
β Reason β
β Next Action β
ββββββββ¬ββββββββ
β
βΌ
ββββββββββββββββ
β Act β
β Click / Type β
β Navigate β
ββββββββ¬ββββββββ
β
βΌ
ββββββββββββββββ
β Repeat β
β Until Done β
ββββββββββββββββ
Intent β Observation β Reasoning β Action β Result
The browser becomes the execution environment, while the agent handles the decision-making.
| Traditional automation | Aevora |
|---|---|
| Step-by-step scripts | Goal-oriented execution |
| Hard-coded selectors | AI-guided interaction |
| Fixed workflows | Adaptive workflows |
| Script-first | Agent-first |
| Manual browser logic | Intelligence + browser control |
Aevora is not just about automating clicks. It is about building systems that can understand what needs to happen and act through a real interface.
Natural-language goals and intelligent decision-making.
Navigate, interact, and execute across web interfaces.
Build repeatable workflows that go beyond rigid scripts.
Aevora sits at the intersection of:
- Artificial Intelligence
- LLM-powered agents
- Browser automation
- Web interaction
- Python
- Agent tools
- Task execution
The project is designed to explore how AI can interact with the web through intelligent, goal-oriented workflows.
git clone https://github.com/ATLURI0001/Aevora.git
cd Aevorapip install -r requirements.txtpython main.pyNote: Update these commands to match the actual project structure as development progresses.
Aevora is an evolving open-source project focused on intelligent browser automation.
The long-term vision is to build a foundation where AI agents can:
- Understand complex web tasks.
- Navigate dynamic interfaces.
- Execute multi-step workflows.
- Extract useful information.
- Integrate browser capabilities into larger AI systems.
The goal is simple: make the web more accessible to intelligent agents.
Contributions are welcome.
If you are interested in AI agents, browser automation, web interaction, or developer tooling, you can help shape Aevora.
# Fork β Branch β Build β Commit β Pull Request- Fork the repository.
- Create a feature branch.
- Make your changes.
- Commit your work.
- Open a pull request.
This project is licensed under the MIT License.
See LICENSE for details.