Heads up, looks like the "*Run Frontend Unit Tests*" job of our CI is flakey since (atleast) yesterday. I'll try to take a look soon, but for now re-running the job if it fails should work

I think the openhands serve commands may be broken?

👋 Hello World 🌎 I'm trying to get OpenHands to work locally.. looking at my local github folder and files.. what's the fastest way to do that?

has renamed the channel from "eng-help" to "dev-help"

Any ideas on why the "Enterprise preview" job fails? https://github.com/OpenHands/OpenHands/pull/11645 cc @quiet-night-61895

I messed with advance configurations in OpenHands cloud and now I can't get it work with OpenHands claude, etc. Where can I find what i need to readjust the settings so that this authenticates correctly fixed

@calm-artist-7206, @straight-sundown-84319: Can you guys give the Org framework PR a good once over: https://github.com/OpenHands/OpenHands/pull/11265 I've done some basic tests on both OSS and SAAS. There will be significant testing done on staging once this is merged.

Hi, is

conversation.astream()

still supported?

async def run_agent():
    conversation = Conversation(agent=agent, workspace=cwd)
    conversation.send_message("Write 3 facts about Python to FACTS.txt")
    
    async for event in conversation.astream():
        print(f"Event: {event}")

This is from the online docs: here. When I try to run this code snippet locally, I get this error:

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
Cell In[2], line 1
----> 1 await run_agent()

Cell In[1], line 40, in run_agent()
     37 conversation = Conversation(agent=agent, workspace=cwd)
     38 conversation.send_message("Please add 'hi there' to the top of the readme.md file")
---> 40 async for event in conversation.astream():
     41     print(f"Event: {event}")

AttributeError: 'LocalConversation' object has no attribute 'astream'

Can we cut down on the size of our root directory? https://github.com/OpenHands/OpenHands/

Gemini 3 Support 🧵

are CI's still building for people? I'm getting the following error for OpenHands repo CIs

Hello everyone, The pipeline is currently failing at “Lint / Lint Python (pull_request)”. Please refer to the image below for additional details. Has anyone else experienced this issue? I would appreciate any suggestions or insights on how to resolve it. Thank you! thankyou

hey everyone, i would like to discuss the previously closed thread https://github.com/OpenHands/OpenHands-CLI/issues/82 • i'm on latest release cli-1.3.0, and i still see restricted schema being sent through liteLLM -> openAI.

@strong-ice-84145 regarding the eval workflow: It seems to override the

metadata.json

created by the

benchmarks

repo, is this intended? 👀

Could someone here help me setup LLM in open-hands? I've run it on server in docker container, everything seems to work until except LLM. I'm trying to use codestral API key with chat endpoint they gave me but it doesn't work. I've tried it with advanced settings and their link and open-hand preset on codestral-latest, both returned same error

I thought the “failed to deserialize response” error in ACP should be fixed, right? I just got it now (cc @chilly-france-44439 @quick-window-91705)

This message was deleted.

This message was deleted.

@limited-student-10747 it seems a recent PR for CLI is failing lint and tests, mind taking a look whenever you have time? (i haven't blocked pr merges based on failing tests, so apologies for the inconvenience 🙏)

Hello everyone, how's the day treating you? New here

@elegant-branch-54423 question regarding the flag

--output json

in https://github.com/OpenHands/OpenHands-CLI/issues/147 Is this a type of headless mode but instead of a rich TUI we output json? So

--headless --output json

is the same

--output json

Or should

--headless

mean we never emit any events?

Hello everyone, I’m new here. I’ve noticed that openHands leverages locAgent, yet it doesn’t expose locAgent’s capabilities to CodeActAgent—preventing CodeActAgent from utilizing the repository comprehension feature while writing code. This means that if I want CodeActAgent to access locAgent’s functionality, I need to register locAgent’s tool in CodeActAgent’s tool list. Is that correct?

Hello everyone, I'm not sure if I'm posting in the right channel... I'm new to using OpenHands, and I'm encountering a problem when I try to create a conversation: Server error '500 Internal Server Error' for url 'http://host.docker.internal:38693/api/conversations' For more information check: https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/500 I installed it via Docker using the official documentation. In the logs: Traceback (most recent call last): 2025-12-18T135705.150362997Z File "/app/openhands/app_server/app_conversation/live_status_app_conversation_service.py", line 266, in _start_app_conversation 2025-12-18T135705.150365223Z response.raise_for_status() 2025-12-18T135705.150366700Z ~~~~~~~~~~~~~~~~~~~~~~~~^^ 2025-12-18T135705.150368057Z File "/app/.venv/lib/python3.13/site-packages/httpx/_models.py", line 829, in raise_for_status 2025-12-18T135705.150369930Z raise HTTPStatusError(message, request=request, response=self) 2025-12-18T135705.150371770Z httpx.HTTPStatusError: Server error '500 Internal Server Error' for url 'http://host.docker.internal:38693/api/conversations' 2025-12-18T135705.150373332Z For more information check: https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/500 In the container agent logs: {"asctime": "2025-12-18 135705,147", "levelname": "INFO", "name": "uvicorn.access", "filename": "h11_impl.py", "lineno": 473, "message": "172.17.0.1:44500 - \"POST /api/conversations HTTP/1.1\" 500"} I'm using ollama with mistral:7b-instruct Thank you for your help.

Build & evaluate workflow seems to be failing on SDK (cc @strong-ice-84145) https://github.com/OpenHands/software-agent-sdk/actions/runs/20464297297/job/58804118238

Getting an error on app.all-hands.dev when starting a conversation

litellm.AuthenticationError: AuthenticationError: Litellm_proxyException - Authentication Error, Invalid proxy server token passed. Received API Key = sk-...zk9w, Key Hash (Token) =****. Unable to find token in cache or `LiteLLM_VerificationTokenTable`

Error on app-all-hands.dev using opus 4.5. Stopped the conversation but it could continue after another user message

Error on app-all-hands.dev with long conversations. When I try reconnecting to a conv I get "An error occurred. Please try again" and I can't log in the conversation anymore.

Hello team , I’m trying to set up the enterprise server locally and could use some guidance. I was advised by @millions-afternoon-47582 to follow the enterprise docs, but I believe they’re slightly outdated: https://github.com/OpenHands/OpenHands/tree/main/enterprise/enterprise_local What I’ve done so far I was able to figure out most of the setup on my own, but I ran into issues at this step: 1.

decrypt_env.sh

step The docs mention running

local/decrypt_env.sh

, but they don’t specify that a path argument is required. From the script itself:

enterprise_local/decrypt_env.sh <DEPLOY_DIR>
Example: enterprise_local/decrypt_env.sh /path/to/deploy

Since this wasn’t mentioned in the docs and I wasn’t sure what path to pass, I skipped this step. 2. Error while running

convert_to_env.py

After that, I ran:

python enterprise_local/convert_to_env.py

And got the following error:

Error: [Errno 2] No such file or directory: 'github_decrypted.yaml'

Traceback (most recent call last):
  File ".../enterprise_local/convert_to_env.py", line 87, in <module>
    os.remove(yaml_file)
FileNotFoundError: [Errno 2] No such file or directory: 'github_decrypted.yaml'

I assume this file should have been generated by the skipped

decrypt_env.sh

step, but I’m not fully sure. Context about my PR I’ve opened a PR here: https://github.com/OpenHands/OpenHands/pull/12181 I tested everything up to the point where clicking the GitHub auth button redirects to GitHub. However, after the redirect, the server was still running in OSS mode, so I couldn’t complete the OAuth flow end-to-end and fully validate the PR. My questions What exactly am I doing wrong in the enterprise local setup? 1. What should the correct

DEPLOY_DIR

be for

decrypt_env.sh

? 2. Is there a recommended way to fully run the enterprise server locally so GitHub auth can be tested end-to-end? 3. Is there anything else I should configure before testing OAuth for enterprise mode? Any help or pointers would be really appreciated. Thanks a lot!

We're archiving this channel, please ask questions on the appropriate "proj-*" channel, or on "questions" if you're not sure what channel to use.

archived the channel