REDEFINING BROWSER INTELLIGENCE

Delivering the world's most advanced simulation and data engine for AI web agents.

Built By Experts From

SCROLL

CAPABILITIES

Foundational pillars for browser intelligence.

SIMULATION

Pixel-perfect, reproducible browser environments—no drift, no noise, no rate limits.

EVALUATION

Every agent action is tracked, classified, and tagged—click failures, layout shifts, and misfires are never silent.

DATA

Expert annotators generate custom, long-horizon datasets for supervised fine-tuning of browser agents on real enterprise platforms.

RL

Safely sample and evaluate unlimited trajectories—train browser agents at scale, without anti-bot constraints.

DEMO

See Foundry in Action

WATCH DEMO

START BUILDING

Integrate Foundry into your workflow and begin building better browser agents.

INTEGRATION

Plug Foundry into your existing agent workflow with our Python SDK.

evaluation_loop.py

for task_id in tasks:
    env = AWE(
      api_key=foundry_api_key, 
      task_id=task_id, 
      record_events=True
    )

    task = await env.get_task()
    await env.reset()
    cdp_url = await env.get_cdp_url()
    
    # Run the agent with run_st_async integration
    result = await agent(
        task.prompt,
        cdp_url,
        task.start_url,
        task.login_credentials,
        task.payment_details,
        task.additional_instructions,
    )

    # get the result, which is the final state of the environment
    final_state = await env.get_state()

    mutations = await env.get_state_mutations()
    events = await env.get_events()

    # compare the result with the ground truth
    score = heuristic(final_state, task.ground_truth)

GET STARTED

Choose your path to begin building with Foundry.

COMPARE RESULTS

GET FULL ACCESS

Ready to Build Better Agents?

Join our private beta and unlock all platform features

SAMPLE DATASET

Coming Soon!

Get sample evaluation data to test locally

5min

Setup Time

10x

Faster Eval