Our autonomous Web Agent is accessible conveniently via an API. The agent empowers you to automate complex Web workflows with the intelligence of cutting-edge techniques.
Start off immediately with our free tier. Paid version with unlimited and parallel calls per API key coming soon.
This API allows you to build applications that can intelligently navigate websites, extract information, interact with online forms, and complete tasks autonomously.
Our core set of technologies involving hierarchical planning, self-refinement and self-improvement lead our Web agent to be the best in its class. READ our technical paper here.
Collect and process data with high accuracy and speed from multiple sources with Web interfaces.
Automate repetitive tasks, allowing your users to focus on more critical activities.
Create automations that can perform continuous testing on web applications to identify bugs, check functionality, or validate user flows.
Discover new ways to automate and optimize your digital tasks.
Shell:
curl -X POST https://api.emergence.ai/orchestrator/v0/agents/web-automation-agent/workflows \
-H "Content-Type: application/json" \
-H "apikey: {your_api_key_here}" \
-d '{
"prompt": "Start with this test prompt"
}'
Achieve high levels of precision in data extraction and task automation. More details of the planning and self-refinement technology and performance evaluation here (link to Agent-E )
The agent’s Hierarchical planning automatically detects and corrects errors, ensuring increased reliability.
The paid version of the API allows secure and parallel execution of tasks within isolated containers for fast results.
Our Pub-Sub architecture allows Asynchronous but necessary Human-Agent communication for monitoring the agent progress and seeking human input for credentials, captchas, permissions to take specific actions, etc.
Python:
import requests
url = "https://api.emergence.ai/orchestrator/v0/agents/web-automation-agent/workflows"
headers = {
"Content-Type": "application/json",
"apikey": "{your_api_key_here}"
}
data = {
"prompt": "You are a QA automation specialist. Please go to Google and test that all buttons and links work. Document any issues found and compile the results into a report"
}
response = requests.post(url, headers=headers, json=data)
print(response.status_code)
print(response.json())