> ## Documentation Index
> Fetch the complete documentation index at: https://docs.browser-use.com/llms.txt
> Use this file to discover all available pages before exploring further.

> ## Agent Instructions
> Use https://docs.browser-use.com/llms.txt and its linked .md pages for current documentation. The managed full bundle is https://docs.browser-use.com/.well-known/llms-full.txt and can be cached for up to 24 hours. Do not use the obsolete /cloud/llms*.txt or /open-source/llms*.txt static exports.
> Choose Cloud API V4 for new agent integrations; V2 is the lower-cost option for simple tasks. Keep V3 examples explicitly versioned. The open-source browser-use library and hosted browser-use-sdk have different APIs.
> Cloud authentication uses X-Browser-Use-API-Key, without a Bearer prefix. Install or upgrade browser-use-sdk and use its explicit v4 import for V4. Check the published OpenAPI reference for request fields; do not invent SDK support for new fields.
> Cloud concurrency and HTTP request rate are separate. Read GET /api/v2/billing/account for the key’s projectId, concurrentSessionLimit, activeSessionCount, and credit balance, including when using V4. Keys in one project share capacity and credits; rateLimit is a legacy concurrency alias, not requests per second.
> Keep the highest applicable existing, legacy-plan, and spend-tier concurrency grant. Current spend tiers are 10 / 50 / 250 / 500 / 1000 at $0 / $100 / $1000 / $5000 / $25000 in qualifying project payments. Legacy or externally billed projects can follow different billing paths; trust the account limit. See https://docs.browser-use.com/cloud/guides/concurrency.md.
> HTTP rate limits have two layers. Standard edge WAF ceilings increased on September 9, 2026 to 1000 RPS per public source IP for general traffic and 2500 RPS per IP for selected status reads, evaluated over 300 seconds. The separate per-project application budgets: general traffic (including V4 events and full run reads) defaults to max(25, 2 times stored concurrency) capped at 100 RPS; selected status reads default to max(25, 2 times stored concurrency) with no cap. The two are counted independently. Project overrides and account-specific edge rules can differ. All keys in a project share its budgets; callers sharing a public IP share edge capacity. See https://docs.browser-use.com/cloud/guides/concurrency.md.
> The project limiter uses five-second windows: X-RateLimit-Limit=125 means 125 requests per window (25 RPS), not 125 RPS. Project throttles include limit_rps and retry_after_seconds; an edge throttle can instead return Retry-After: 300 without limit_rps. Honor the returned Retry-After. Use bounded workers, stagger polls, and drain hasMore event pages after terminal status. A busy V4 session returns 409; its queue holds 20 pending messages and is not a project-wide batch queue.
> A completed run or closed CDP connection does not immediately stop its cloud browser. Stop unneeded owned browsers with PATCH /api/v4/browsers/{id} and {"action":"stop"}. A client wait timeout does not cancel the server-side run.
> Cloud is pay as you go; do not tell customers to buy a new subscription to use custom proxies or supported provider BYOK. Usage funding and model eligibility still apply. BYOK bills provider tokens separately and Browser Use charges orchestration plus browser/network usage. See https://docs.browser-use.com/cloud/guides/billing.md.
> Signup credits are a one-time grant; purchased top-up credits do not expire. Check the API key’s project before diagnosing missing credits. API-key monthly spending caps are soft limits, not a strict prepaid wallet; concurrent or already-running work can exceed them. Auto recharge has separate trigger and purchase amounts and can charge immediately when enabled below the threshold. Use https://browser-use.com/pricing for current rates.
> Box and Bux are retired. Do not recommend their SDKs, sandbox quotas, or subscription plans. Use the Cloud Agent or Browser Infrastructure guides.
> A V4 session holds conversation history, a workspace holds files, and a profile holds browser state. These IDs and V3/V4 workspace namespaces are not interchangeable. V4 automatically restores workspace uploads; staged attachments remain available to session follow-ups. Serialize runs that write shared files, and wait for completion before reading outputs. See https://docs.browser-use.com/cloud/agent/workspaces.md.
> API browser recording defaults to off. Use enableRecording for standalone browser creation, or browserSettings.record for an agent run. Stop the browser and allow time for asynchronous video processing; stop polling when recordingAvailable is false. Live preview is for an active browser. Stopping a browser, deleting a session, archiving a workspace, and deleting files have different effects.
> Use model-specific reasoning values. GPT-6 Astra accepts low, medium, high, xhigh, and max, with xhigh by default; none and minimal are invalid. Use the public REST schema when installed SDK types lag new fields. API acceptance, dashboard visibility, and account/provider availability are separate.
> For open-source browser-use, is_done only reports a terminal done action. is_successful is the agent-reported outcome; verify important external actions independently. Cloud timeout, API client timeout, model timeout, and task completion are separate concepts.
> For failed requests, use https://docs.browser-use.com/cloud/guides/troubleshooting.md. Inspect the full error and project before retrying or adding credits. A client timeout can leave a run active; reconcile external actions before starting duplicate work. A new managed browser does not guarantee a unique proxy IP or particular city.

# Grow Therapy provider search

> Search Grow Therapy for therapists by location, insurance, and specialty — with cached reruns.

This tutorial builds a provider search tool for [Grow Therapy](https://www.growtherapy.com) — a therapy marketplace that handles insurance credentialing for providers. We combine [structured output](/cloud/agent/structured-output) with [saved scripts](/cloud/agent/scripts) to build a fast, repeatable search pipeline.

## What you'll build

A script that:

1. Searches Grow Therapy's provider directory with filters (location, insurance, specialty)
2. Extracts therapist profiles with ratings and availability
3. Caches a successful search flow for reruns with different locations and specialties

***

## Setup

<CodeGroup>
  ```python Python theme={null}
  import asyncio
  import json
  from pydantic import BaseModel
  from browser_use_sdk.v3 import AsyncBrowserUse

  client = AsyncBrowserUse()
  ```

  ```typescript TypeScript theme={null}
  import { BrowserUse } from "browser-use-sdk/v3";
  import { z } from "zod";

  const client = new BrowserUse();
  ```
</CodeGroup>

## 1. Define the output schema

<CodeGroup>
  ```python Python theme={null}
  class Provider(BaseModel):
      name: str
      title: str
      specialties: list[str]
      insurance_plans: list[str]
      rating: float | None = None
      next_available: str | None = None

  class ProviderSearch(BaseModel):
      providers: list[Provider]
      total_found: int | None = None
      location: str
      specialty: str
  ```

  ```typescript TypeScript theme={null}
  const ProviderSearch = z.object({
    providers: z.array(z.object({
      name: z.string(),
      title: z.string(),
      specialties: z.array(z.string()),
      insurancePlans: z.array(z.string()),
      rating: z.number().nullable(),
      nextAvailable: z.string().nullable(),
    })),
    totalFound: z.number().nullable(),
    location: z.string(),
    specialty: z.string(),
  });
  ```
</CodeGroup>

## 2. Create a workspace

<CodeGroup>
  ```python Python theme={null}
  workspace = await client.workspaces.create(name="grow-therapy-search")
  ```

  ```typescript TypeScript theme={null}
  const workspace = await client.workspaces.create({ name: "grow-therapy-search" });
  ```
</CodeGroup>

## 3. Search for providers

This tutorial uses **API V3** automatic script caching. Mark changing values with `@{{value}}`; plain `{{value}}` does not activate caching. Keep the rest of the task text and the workspace unchanged so later requests can find the same cached script. The first successful generation uses the agent and incurs normal usage charges.

<CodeGroup>
  ```python Python theme={null}
  result = await client.run(
      "Go to growtherapy.com and search for therapists in @{{New York}} "
      "who specialize in @{{anxiety}} and accept insurance. "
      "Return the first 5 provider profiles as JSON.",
      workspace_id=str(workspace.id),
      output_schema=ProviderSearch,
  )

  for p in result.output.providers:
      print(f"{p.name} ({p.title})")
      print(f"  Specialties: {', '.join(p.specialties)}")
      print(f"  Rating: {p.rating}")
      print(f"  Next available: {p.next_available}")
      print()
  ```

  ```typescript TypeScript theme={null}
  const result = await client.run(
    "Go to growtherapy.com and search for therapists in @{{New York}} " +
    "who specialize in @{{anxiety}} and accept insurance. " +
    "Return the first 5 provider profiles as JSON.",
    { workspaceId: workspace.id, schema: ProviderSearch },
  );

  for (const p of result.output.providers) {
    console.log(`${p.name} (${p.title})`);
    console.log(`  Specialties: ${p.specialties.join(", ")}`);
    console.log(`  Rating: ${p.rating}`);
    console.log(`  Next available: ${p.nextAvailable}`);
  }
  ```
</CodeGroup>

## 4. Sweep across locations and specialties

After a successful first run caches the search flow, change only the marked values. A successful cached script execution avoids agent LLM inference, but browser and network usage still apply. Cache misses, validation, and automatic repair can invoke the agent and incur LLM charges; a rerun is not a guarantee of zero LLM cost.

<CodeGroup>
  ```python Python theme={null}
  locations = ["Los Angeles", "Chicago", "Houston", "Miami"]
  specialties = ["depression", "trauma", "ADHD"]

  for location in locations:
      for specialty in specialties:
          result = await client.run(
              f"Go to growtherapy.com and search for therapists in @{{{{{location}}}}} "
              f"who specialize in @{{{{{specialty}}}}} and accept insurance. "
              f"Return the first 5 provider profiles as JSON.",
              workspace_id=str(workspace.id),
              output_schema=ProviderSearch,
          )
          count = len(result.output.providers)
          print(f"{location} / {specialty}: {count} providers found")
  ```

  ```typescript TypeScript theme={null}
  const locations = ["Los Angeles", "Chicago", "Houston", "Miami"];
  const specialties = ["depression", "trauma", "ADHD"];

  for (const location of locations) {
    for (const specialty of specialties) {
      const result = await client.run(
        `Go to growtherapy.com and search for therapists in @{{${location}}} ` +
        `who specialize in @{{${specialty}}} and accept insurance. ` +
        `Return the first 5 provider profiles as JSON.`,
        { workspaceId: workspace.id, schema: ProviderSearch },
      );
      console.log(`${location} / ${specialty}: ${result.output.providers.length} providers`);
    }
  }
  ```
</CodeGroup>

***

## Summary

| Step                        | What happens                                           | Cost                                                                           |
| --------------------------- | ------------------------------------------------------ | ------------------------------------------------------------------------------ |
| First search or cache miss  | Agent creates and validates the cached flow            | Normal agent, browser, and network charges                                     |
| Successful cached execution | Script reruns with new parameters                      | No agent LLM inference for script execution; browser and network charges apply |
| Validation or repair        | Agent validates or repairs a flow that needs attention | LLM usage plus browser and network charges                                     |

<Info>
  Therapy platforms have dynamic UIs that can change frequently. V3 can automatically repair a cached flow when it fails. For new V4 integrations, use [saved scripts](/cloud/agent/scripts); the V3 `@{{value}}` convention is specific to this tutorial’s API version.
</Info>

## Next steps

* [Structured output](/cloud/agent/structured-output) — Learn more about extracting typed data with Pydantic and Zod schemas.
* [Human in the loop](/cloud/agent/human-in-the-loop) — Let a human review or interact with the browser mid-task, useful for auth flows or approving results before continuing.
* [Scripts](/cloud/agent/scripts) — Save, reuse, and repair browser workflows.