Zeeland

The calling time of function pezzo.get_prompt() is too long

1

### Proposal Here is my code: ```python from pezzo.client import pezzo import time start_time = time.time() res = pezzo.get_prompt("001") duration = time.time() - start_time print(duration) res = pezzo.get_prompt("001") duration =...

## 🚀 Feature Request ref: https://github.com/jorgearango/llmapper An experiment in using LLMs to draw simple concept maps. Currently, a simple bash script that uses Simon Willison's [llm](https://github.com/simonw/llm), [strip-tags](https://github.com/simonw/strip-tags), and [ttok](https://github.com/simonw/ttok) tools...

Add gradio+ pne.chat simple chat app use cases

1

## 🚀 Feature Request Just build a simple chatbot by gradio and pne.chat Functions: - select models - input env and key ## Output - a example code in examples...

## 🚀 Feature Request Just build a simple chatbot by streamlit and pne.chat Functions: - select models - input env and key ## Output - a example code in examples...

Remove useless packages when pip it

1

### 🐛 Describe the bug I have seen that poetry configures the dependency of several models, I believe that the user will only use one model when using, hope to...

## 🚀 Feature Request 在一些场景里，我们需要用合适的模型去应用在不同的场景里，比如在简单问答的场景下，可以直接的使用 gpt-3.5，在需要复杂逻辑推理的场景下，用 gpt-4-turbo。 这意味着，对于一个问题来说，需要一种路由模型，来选择出一个合适的推理模型，下面还一种可能的示例代码。 ```python import pne pne.chat(messages="hello", model=["gpt-4o", "deepseek/deepseek-chat"]) ``` ref: https://x.com/tuturetom/status/1819400584799441162 ## Question - 对于 [https://x.com/tuturetom/status/1819400584799441162](https://x.com/tuturetom/status/1819400584799441162) 中的代码演示，或许我们还需要判断一下，或许这种调用方式并不是最合适的调用方式，待讨论。 - 在很多场景下，用一些简单的逻辑判断就可以构建出 model router 的功能，model router 这个比较强场景相关。...

## 🚀 Feature Request Add a example to show how to use e2b code interpreter in ToolAgent + streamlit. ref: https://github.com/e2b-dev/e2b-cookbook/blob/main/examples/claude-code-interpreter-python/claude_code_interpreter.ipynb example ref: https://www.promptulate.cn/#/use_cases/build-math-application-with-agent?id=building-a-math-application-with-promptulate-agents

## 🚀 Feature Request Currently, if you use tool in pne.chat. ReAct prompt pattern will turn on default. Now lots of model support tool call, it more simple and fast....

Add stream output for Agent

1

Related to #833 Add stream output support for ToolAgent. * **ToolAgent Class**: - Add `stream` parameter to the `run` method. - Implement logic to handle the `stream` parameter. - Modify...

Add stream output for Agent

1

## 🚀 Feature Request Current, ToolAgent cannot use stream mode output. As a ReAct prompt pattern, the intermediate thought steps need to be output. Now we use hooks system to...