midscene

midscene copied to clipboard

feat: invoke anthropic SDK to call Claude

1

yuyutaotao

run demo failed, Error: 400 Missing required parameter: 'response_format.json_schema'. (request id: 202412201428294827994785579509)

9

/Users/hankunfang/workspace/midscene-example/puppeteer-demo/node_modules/@midscene/web/dist/lib/puppeteer.js:1291 throw new Error(`${errorTask == null ? void 0 : errorTask.error} ^ Error: 400 Missing required parameter: 'response_format.json_schema'. (request id: 202412201428294827994785579509) Error: 400 Missing required parameter: 'response_format.json_schema'. (request id: 202412201428294827994785579509)...

hankunfang

How to implement a a 'goal-to-task' AI planning tool based on Midscene

3

[The FAQ](https://midscenejs.com/faq.html#can-midscene-smartly-plan-the-actions-according-to-my-one-line-goal-like-executing-tweet-hello-world) mentioned > If you require a 'goal-to-task' AI planning tool, you can develop one based on Midscene. But I'm not sure how to leverage Midscene to make this,...

ducan-ne

Support for elements not visible In current viewport

2

Midscene is currently limited to interacting with the current viewport, but in real-world scenarios, we need to deal with elements not visible in the current viewport (e.g., elements that need...

ducan-ne

执行操作异常

1

Failed to plan actions with context: undefined AssertionError [ERR_ASSERTION]: Failed to plan actions with context: undefined at Sr (chrome-extension://gbldofcpkknbggpkmbdaefngejllnief/lib/popup.js:1:2816013) at async Object.executor (chrome-extension://gbldofcpkknbggpkmbdaefngejllnief/lib/popup.js:1:2873406) at async Vt.flush (chrome-extension://gbldofcpkknbggpkmbdaefngejllnief/lib/popup.js:1:2790577) at async N.action...

zjian95

fix: fix scroll action in chrome extension

1

yuyutaotao

parse json error: ```json

1

使用 qwen 配置为： OPENAI_API_KEY sk-xxxxxxxxxxx OPENAI_BASE_URL https://dashscope.aliyuncs.com/compatible-mode/v1 MIDSCENE_MODEL_NAME qwen-vl-max-lates 运行 plawright-demo e2e 文件夹中的 “todo-mvc-zh.spec.ts” 文件，即： import { expect } from "@playwright/test"; import { test } from "./fixture"; test.beforeEach(async ({ page...

wenmhan

Can we support the content within iframe?

2

I tried the page with the iframe and found that its image analysis could not cover the contents of the iframe, as shown in the following figure:   I'm...

tdhanjun

[Feature] llm Options

1

# Background I wish to experiment with other LLM models like Tongyi or Claude, but it's not feasible. https://github.com/web-infra-dev/midscene/blob/863e5a54d0380004038b6012d4810d79a437a401/packages/midscene/src/ai-model/openai/index.ts#L94 # Aspiration Provide a solution akin to [browser-use](https://github.com/gregpr07/browser-use), allowing the integration...

729993031

yuyutaotao