chore: rip perform

docs/src/api/class-page.md

@@ -710,54 +710,6 @@ current working directory.
 
 Raw CSS content to be injected into frame.
 
-## async method: Page.agent
-* since: v1.58
-* langs: js
-- returns: <[PageAgent]>
-
-Initialize page agent with the llm provider and cache.
-
-### option: Page.agent.cache
-* since: v1.58
-- `cache` <[Object]>
-  - `cacheFile` ?<[string]> Cache file to use/generate code for performed actions into. Cache is not used if not specified (default).
-  - `cacheOutFile` ?<[string]> When specified, generated entries are written into the `cacheOutFile` instead of updating the `cacheFile`.
-
-### option: Page.agent.expect
-* since: v1.58
-- `expect` <[Object]>
-  - `timeout` ?<[int]> Default timeout for expect calls in milliseconds, defaults to 5000ms.
-
-### option: Page.agent.limits
-* since: v1.58
-- `limits` <[Object]>
-  - `maxTokens` ?<[int]> Maximum number of tokens to consume. The agentic loop will stop after input + output tokens exceed this value. Defaults to unlimited.
-  - `maxActions` ?<[int]> Maximum number of agentic actions to generate, defaults to 10.
-  - `maxActionRetries` ?<[int]> Maximum number retries per action, defaults to 3.
-
-Limits to use for the agentic loop.
-
-### option: Page.agent.provider
-* since: v1.58
-- `provider` <[Object]>
-  - `api` <[PageAgentAPI]<"openai"|"openai-compatible"|"anthropic"|"google">> API to use.
-  - `apiEndpoint` ?<[string]> Endpoint to use if different from default.
-  - `apiKey` <[string]> API key for the LLM provider.
-  - `apiTimeout` ?<[int]> Amount of time to wait for the provider to respond to each request.
-  - `model` <[string]> Model identifier within the provider. Required in non-cache mode.
-
-### option: Page.agent.secrets
-* since: v1.58
-- `secrets` ?<[Object]<[string], [string]>>
-
-Secrets to hide from the LLM.
-
-### option: Page.agent.systemPrompt
-* since: v1.58
-- `systemPrompt` <[string]>
-
-System prompt for the agent's loop.
-
 ## async method: Page.bringToFront
 * since: v1.8
 

docs/src/test-api/class-fixtures.md

@@ -18,10 +18,6 @@ Given the test above, Playwright Test will set up the `page` fixture before runn
 
 Playwright Test comes with builtin fixtures listed below, and you can add your own fixtures as well. Playwright Test also [provides options][TestOptions] to  configure [`property: Fixtures.browser`], [`property: Fixtures.context`] and [`property: Fixtures.page`].
 
-## property: Fixtures.agent
-* since: v1.58
-- type: <[PageAgent]>
-
 ## property: Fixtures.browser
 * since: v1.10
 - type: <[Browser]>

docs/src/test-api/class-fullconfig.md

@@ -104,15 +104,6 @@ See [`property: TestConfig.reportSlowTests`].
 
 Base directory for all relative paths used in the reporters.
 
-## property: FullConfig.runAgents
-* since: v1.58
-- type: <['RunAgentsMode]<"all"|"missing"|"none">>
-
-Whether to run LLM agent for [PageAgent]:
-* "all" disregards existing cache and performs all actions via LLM
-* "missing" only performs actions that don't have generated cache actions
-* "none" does not talk to LLM at all, relies on the cached actions (default)
-
 ## property: FullConfig.shard
 * since: v1.10
 - type: <[null]|[Object]>

docs/src/test-api/class-testconfig.md

@@ -515,15 +515,6 @@ export default defineConfig({
 });
 ```
 
-## property: TestConfig.runAgents
-* since: v1.58
-- type: ?<['RunAgentsMode]<"all"|"missing"|"none">>
-
-Whether to run LLM agent for [PageAgent]:
-* "all" disregards existing cache and performs all actions via LLM
-* "missing" only performs actions that don't have generated cache actions
-* "none" does not talk to LLM at all, relies on the cached actions (default)
-
 ## property: TestConfig.shard
 * since: v1.10
 - type: ?<[null]|[Object]>

docs/src/test-api/class-testoptions.md

@@ -46,23 +46,6 @@ export default defineConfig({
 });
 ```
 
-## property: TestOptions.agentOptions
-* since: v1.58
-- type: <[Object]>
-  - `provider` <[Object]>
-    - `api` <[PageAgentAPI]<"openai"|"openai-compatible"|"anthropic"|"google">> API to use.
-    - `apiEndpoint` ?<[string]> Endpoint to use if different from default.
-    - `apiKey` <[string]> API key for the LLM provider.
-    - `apiTimeout` ?<[int]> Amount of time to wait for the provider to respond to each request.
-    - `model` <[string]> Model identifier within the provider. Required in non-cache mode.
-  - `cachePathTemplate` ?<[string]> Cache file template to use/generate code for performed actions into.
-  - `limits` <[Object]>
-    - `maxTokens` ?<[int]> Maximum number of tokens to consume. The agentic loop will stop after input + output tokens exceed this value. Defaults to unlimited.
-    - `maxActions` ?<[int]> Maximum number of agentic actions to generate, defaults to 10.
-    - `maxActionRetries` ?<[int]> Maximum number retries per action, defaults to 3.
-  - `secrets` ?<[Object]<[string], [string]>> Secrets to hide from the LLM.
-  - `systemPrompt` <[string]> System prompt for the agent's loop.
-
 ## property: TestOptions.baseURL = %%-context-option-baseURL-%%
 * since: v1.10
 

examples/todomvc/tests/fixtures.ts

@@ -5,14 +5,6 @@ import { test as baseTest } from '@playwright/test';
 export { expect } from '@playwright/test';
 
 export const test = baseTest.extend({
-  agentOptions: {
-    provider: {
-      api: 'anthropic',
-      apiKey: process.env.AZURE_SONNET_API_KEY!,
-      apiEndpoint: process.env.AZURE_SONNET_ENDPOINT!,
-      model: 'claude-sonnet-4-5',
-    },
-  },
   page: async ({ page }, use) => {
     await page.goto('https://demo.playwright.dev/todomvc');
     await use(page);