langgraph/how-tos/local-studio/ #2472
giscus[bot]
bot
Replies: 18 comments 41 replies
-
|
I keep getting some starlette error after running Traceback (most recent call last): |
Beta Was this translation helpful? Give feedback.
-
|
It should all work via the langgraph SDK/API! You'd just set the url as https://langchain-ai.github.io/langgraph/cloud/deployment/test_locally/ |
Beta Was this translation helpful? Give feedback.
-
|
ah okay, so I have to use a langgraph api server first (been using my own fastapi wrapper on top). okay will go through the server some other time! |
Beta Was this translation helpful? Give feedback.
-
|
This is a LangGraph API server! |
Beta Was this translation helpful? Give feedback.
-
|
right but I don't have some sort of opentelemetry way of just saying 'hey start tracking when this function is called', it has to be a call to the server. don't have time right now to refactor such that my fastapi calls the langgraph api server instead of just graph.invoke in the code right now lol |
Beta Was this translation helpful? Give feedback.
-
|
Hm I'm not sure I fully understand so let me know if you're blocked. The code all can be traced with langsmith if you set LANGSMITH_TRACING=true, and you can trace nested fucntion call with the |
Beta Was this translation helpful? Give feedback.
-
|
Trying to install it i get this: Wasnt the point of this to have a similar UI as the studio but in another OS? or did i misunderstood it? |
Beta Was this translation helpful? Give feedback.
-
|
@jabrPromtior could you upgrade please and let me know if you still run into any difficulties? I had inadvertently released the initial version using uvicorn's standard deps, which aren't supported by windows. |
Beta Was this translation helpful? Give feedback.
-
|
Sorry forgot to respond, it was that yes, worked perfectly, thank you! |
Beta Was this translation helpful? Give feedback.
-
|
Do we need to provide a LangSmith api key on startup, as was the case for the dockerized langgraph-api setup? |
Beta Was this translation helpful? Give feedback.
-
|
No |
Beta Was this translation helpful? Give feedback.
-
Beta Was this translation helpful? Give feedback.
-
|
@luizweb did you install your package's dependencies? What errors were shown in the server logs? |
Beta Was this translation helpful? Give feedback.
-
|
How can I pass in a runnable configuration when using the local development server? The configuration option is not available in the online version as in the local studio app. |
Beta Was this translation helpful? Give feedback.
-
|
What are you looking to configure? You should still be able to provide the same metadata information, as the frontend code is identical. To show which configurable values are available, you'd pass in a config schema (https://langchain-ai.github.io/langgraph/cloud/reference/sdk/python_sdk_ref/?h=config+schema#langgraph_sdk.schema.GraphSchema.config_schema) |
Beta Was this translation helpful? Give feedback.
-
|
I've just installed LangGraph locally, using "pip install "langgraph-cli[inmem]==0.1.55". When i try to run it using "langgraph dev" I get the message "Error: Invalid value for '--config': Path 'langgraph.json' does not exist." Where do I get the 'langgraph.json'? Is it supposed to install a default file on installation, or do I need to create this file? |
Beta Was this translation helpful? Give feedback.
-
|
You need to set it up for Langgraph Cloud as per this guide https://langchain-ai.github.io/langgraph/concepts/application_structure/ |
Beta Was this translation helpful? Give feedback.
-
|
Do I have to follow that structure? I get a {
"dependencies": ["."],
"graphs": {
"explore_team": "teams/explore_agent/graph.py:explore_graph",
"compute_team": "teams/compute_agent/graph.py:compute_graph",
"agent": "teams/graph.py:graph"
},
"env": ".env"
}I use the following structure |
Beta Was this translation helpful? Give feedback.
-
|
I have the same issue |
Beta Was this translation helpful? Give feedback.
-
|
try to add the current directory to the Python module search path: |
Beta Was this translation helpful? Give feedback.
-
|
You have to install the current project for it to run or Do you still run into the issue after that? |
Beta Was this translation helpful? Give feedback.
-
|
I fixed it with PYTHONPATH. But would be nice if this was not necessary. |
Beta Was this translation helpful? Give feedback.
-
|
Why does it not load the .env file like this: {
"dependencies": ["."],
"graphs": {
"prometheus": "./backend/graphs/prometheus.py:prometheus"
},
"python-version": "3.11",
"env": ".env"
} |
Beta Was this translation helpful? Give feedback.
-
|
Due to my negligence. Could you try again with and then re-run? |
Beta Was this translation helpful? Give feedback.
-
|
Yeah, it works now. Thanks 🔥 |
Beta Was this translation helpful? Give feedback.
-
|
Failed to load assistants TypeError: Failed to fetch |
Beta Was this translation helpful? Give feedback.
-
|
What browser are you using? There is also a known issue right now where safari does not permit connecting to localhost |
Beta Was this translation helpful? Give feedback.
-
|
I'm using Google Chrome on Linux Ubuntu. With 127.0.0.1 in the address it didn't work. I switched to 'localhost' and it worked. |
Beta Was this translation helpful? Give feedback.
-
|
I'm using Google Chrome on Windows 11. I'll try later. Thank you. |
Beta Was this translation helpful? Give feedback.
-
|
Hm ok - thank you for reporting - I'll get a fix out |
Beta Was this translation helpful? Give feedback.
-
|
Working after using the default UI link AND adding the current directory to the path: On Windows command prompt:
|
Beta Was this translation helpful? Give feedback.
-
|
Would be nice if the --reload-include was available too. |
Beta Was this translation helpful? Give feedback.
-
|
In the context of langgraph dev |
Beta Was this translation helpful? Give feedback.
-
|
Studio loads but I am not able to run workflows, getting "Background run failed" followed by "langgraph.errors.InvalidUpdateError: Expected node messages to update at least one of ['messages'], got {}". Same code works in desktop LG Studio. Any ideas? |
Beta Was this translation helpful? Give feedback.
-
|
Just solved my own problem but worth clarifying expected behavior here anyway -- my workflows do not require human input, so I usually just hit Submit to get them started. That apparently worked in Desktop Studio but doesn't work any more. If I say "start please" as a first message and then hit Submit, it works. Is there a switch I can flip so this doesn't require a message from a human to get started? |
Beta Was this translation helpful? Give feedback.
-
|
One other issue -- cancelling workflows in progress does not appear to actually cancel them -- I am seeing my agents continue to execute in the LangGraph server dev logs. Even stopping the server and restarting it doesn't seem to abort the workflow I cancelled, it just starts again. Is there any way to fix that? |
Beta Was this translation helpful? Give feedback.
-
|
I am running langgraph dev and getting the following error: Environment Details: Any ideas on why this issue occurs specifically with langgraph dev? How can I resolve this? |
Beta Was this translation helpful? Give feedback.
-
|
pip list aiohappyeyeballs 2.4.3 |
Beta Was this translation helpful? Give feedback.
-
|
This is the complete console o/p - ╦ ┌─┐┌┐┌┌─┐╔═╗┬─┐┌─┐┌─┐┬ ┬
This in-memory server is designed for development and testing. 2024-11-22T22:52:21.485583Z [error ] Traceback (most recent call last): 2024-11-22T22:52:21.485814Z [error ] Application startup failed. Exiting. |
Beta Was this translation helpful? Give feedback.
-
|
It's likely that the langgraph CLI path you're using isn't in the current environment and it' referencing a different set of libs |
Beta Was this translation helpful? Give feedback.
-
|
Yes, that was the problem. I uninsalled the global langgraph-cli and then it working on my venv. Thanks |
Beta Was this translation helpful? Give feedback.
-
|
I am unable to make it work. I get |
Beta Was this translation helpful? Give feedback.
-
|
Sounds like your server never started? |
Beta Was this translation helpful? Give feedback.
-
|
Had an error after running langgraph dev It looks like something about Starlette. • Python: 3.11.10 pip list: annotated-types 0.7.0 And here is the log after running Really appreciate if anyone can help 🙏🙏🙏🙏🙏 |
Beta Was this translation helpful? Give feedback.
-
|
It looks like you're trying to use a JS graph? JS isn't supported yet. |
Beta Was this translation helpful? Give feedback.
-
|
Hi there! Is there any ETA on when we can expect this to work w JS graph? My team is working on Open Canvas and would love to reduce friction here! |
Beta Was this translation helpful? Give feedback.
-
|
I try to use local-studio on Ubuntu 22 (WSL) with the following versions: Please help. Thank you. |
Beta Was this translation helpful? Give feedback.
-
|
2 more remarks:
|
Beta Was this translation helpful? Give feedback.
-
|
In-mem API only supports python 3.11 and up, due to some python asyncio features it depends on at present |
Beta Was this translation helpful? Give feedback.
-
|
I have the same problem. I've tried it in windows and linux |
Beta Was this translation helpful? Give feedback.
-
|
Just installed langgraph like you suggested: pip install -U "langgraph-cli[inmem]" python-dotenv then got: langgraph --version This did not work: pip install "langgraph-api-inmem>=0.0.4" Added this to my .bashrc: export PYTHONPATH=$(pwd):$PYTHONPATH My current structure: ~/my-agents$ tree more langgraph.json Very first time testing langgraph, my initial goal is to use it Langgraph Studio server locally with Ollama models... any tips you might provide would be greatly appreciated ! Thanks. |
Beta Was this translation helpful? Give feedback.
-
|
forgot to paste this: pip install -e . |
Beta Was this translation helpful? Give feedback.
-
|
Just made some progress by watching the vid: then using the same "langgraph.json" file showed in the vid. Btw, it would be nice to share the repo being showed in that vid, please ? python3 --version Ubuntu 22.04 LTS langgraph --version langgraph dev ╦ ┌─┐┌┐┌┌─┐╔═╗┬─┐┌─┐┌─┐┬ ┬
This in-memory server is designed for development and testing. 2024-12-03T21:06:54.298638Z [info ] Will watch for changes in these directories: ['/home/alberto/my-agents'] [uvicorn.error] api_variant=local_dev |
Beta Was this translation helpful? Give feedback.
-
|
hello, ModuleNotFoundError: No module named 'my_agent' If you are using pyproject.toml or setuptools: Make sure to run this command from your project's root directory (where your setup.py or pyproject.toml is located) As I used an official example, I guess the structure is correct. Does anyone have an idea how to tackle this issue ? Thank you ! |
Beta Was this translation helpful? Give feedback.
-
|
It's likely a local environment issue where the langgraph CLI version may have been installed globally and then doesn't see your local virtualenv Maybe try using a package manager? E.g., if you wanted to use curl -LsSf https://astral.sh/uv/install.sh | sh
uvx --refresh --from "langgraph-cli[inmem]" --with-editable . --python 3.11 langgraph dev |
Beta Was this translation helpful? Give feedback.
-
langgraph/how-tos/local-studio/
Build language agents as graphs
https://langchain-ai.github.io/langgraph/how-tos/local-studio/
Beta Was this translation helpful? Give feedback.
All reactions