DocsGPT
DocsGPT copied to clipboard
arc53
🐛 Bug Report: Malformed request body error on /stream endpoint
📜 Description
After installing and starting the server via Docker, any request sent from the local UI to the /stream endpoint fails with a "Malformed request body" error on the client side. The backend logs inside the Docker container show that the request data is received but then immediately rejected with the same error. I use the method 4, via OpenAI api key.
👟 Reproduction steps
The Quickstart guide from DocsGPT docs for Windows.
👍 Expected behavior
The /stream endpoint should accept the JSON payload, process the request, and return a streamed response to the client without error.
👎 Actual Behavior with Screenshots
lient: Displays error banner or alert with message
Malformed request body
Backend logs (inside Docker container):
[2025-06-18 10:24:06,808] INFO in routes: /stream - request_data: {
'question': 'What is DocsGPT?',
'history': '[{"prompt":"What is DocsGPT?"}]',
'conversation_id': None,
'prompt_id': 'default',
'chunks': '2',
'token_limit': 2000,
'isNoneDoc': False,
'save_conversation': True,
'retriever': 'duckduck_search'
}, source: {}, attachments: 0
[2025-06-18 10:24:06,808] ERROR in routes: /stream - error: Malformed request body
💻 Operating system
Windows
What browsers are you seeing the problem on?
Firefox, Chrome
🤖 What development environment are you experiencing this bug on?
Docker
🔒 Did you set the correct environment variables in the right path? List the environment variable names (not values please!)
API_KEY LLM_NAME MODEL_NAME VITE_API_STREAMING
📃 Provide any additional context for the Bug.
No response
📖 Relevant log output
Connect Cloud API Provider
Choose your Cloud API Provider:
1) OpenAI
2) Google (Vertex AI, Gemini)
3) Anthropic (Claude)
4) Groq
5) HuggingFace Inference API
6) Azure OpenAI
7) Novita
b) Back to Main Menu
Choose option (1-7, or b): 1
Your API key will be stored locally in the .env file and will not be sent anywhere else
Please enter your API key: sk-...
Configuring for Cloud API Provider: OpenAI...
.env file configured for OpenAI.
Docker is not running. Attempting to start Docker Desktop...
Waiting for Docker to start ...
Docker has started successfully!
Starting Docker Compose...
time="2025-06-18T13:14:44+03:00" level=warning msg="The \"LLM_PROVIDER\" variable is not set. Defaulting to a blank string."
time="2025-06-18T13:14:44+03:00" level=warning msg="The \"OPENAI_BASE_URL\" variable is not set. Defaulting to a blank string."
time="2025-06-18T13:14:44+03:00" level=warning msg="The \"LLM_PROVIDER\" variable is not set. Defaulting to a blank string."
Compose can now delegate builds to bake for better performance.
To do so, set COMPOSE_BAKE=true.
[+] Building 3.8s (41/49) docker:desktop-linux
=> [backend internal] load build definition from Dockerfile 0.0s
=> => transferring dockerfile: 2.67kB 0.0s
=> WARN: FromAsCasing: 'as' and 'FROM' keywords' casing do not match (line 2) 0.0s
=> WARN: FromAsCasing: 'as' and 'FROM' keywords' casing do not match (line 46) 0.0s
=> [worker internal] load build definition from Dockerfile 0.0s
=> => transferring dockerfile: 2.67kB 0.0s
=> WARN: FromAsCasing: 'as' and 'FROM' keywords' casing do not match (line 2) 0.0s
=> WARN: FromAsCasing: 'as' and 'FROM' keywords' casing do not match (line 46) 0.0s
=> [backend internal] load metadata for docker.io/library/ubuntu:24.04 1.3s
=> [worker auth] library/ubuntu:pull token for registry-1.docker.io 0.0s
=> [worker internal] load .dockerignore 0.0s
=> => transferring context: 2B 0.0s
=> [backend internal] load .dockerignore 0.0s
=> => transferring context: 2B 0.0s
=> [worker internal] load build context 0.1s
=> => transferring context: 515.32kB 0.1s
=> [backend builder 1/9] FROM docker.io/library/ubuntu:24.04@sha256:b59d21599a2b151e23eea5f6602f4af4d7d31c4e236d 0.1s
=> => resolve docker.io/library/ubuntu:24.04@sha256:b59d21599a2b151e23eea5f6602f4af4d7d31c4e236d22bf0b62b86d2e38 0.0s
=> [backend internal] load build context 0.1s
=> => transferring context: 515.32kB 0.1s
=> CACHED [worker final 2/9] RUN apt-get update && apt-get install -y software-properties-common && add- 0.0s
=> CACHED [worker final 3/9] WORKDIR /app 0.0s
=> CACHED [worker final 4/9] RUN groupadd -r appuser && useradd -r -g appuser -d /app -s /sbin/nologin -c "D 0.0s
=> CACHED [worker builder 2/9] RUN apt-get update && apt-get install -y software-properties-common && ad 0.0s
=> CACHED [worker builder 3/9] RUN if [ -f /usr/bin/python3.12 ]; then ln -s /usr/bin/python3.12 /usr/bi 0.0s
=> CACHED [worker builder 4/9] RUN wget https://d3dg1063dc54p9.cloudfront.net/models/embeddings/mpnet-base-v2.zi 0.0s
=> CACHED [worker builder 5/9] RUN wget -q -O - https://sh.rustup.rs | sh -s -- -y 0.0s
=> CACHED [worker builder 6/9] RUN apt-get remove --purge -y wget unzip && apt-get autoremove -y && rm -rf /var/ 0.0s
=> CACHED [worker builder 7/9] COPY requirements.txt . 0.0s
=> CACHED [worker builder 8/9] RUN python3.12 -m venv /venv 0.0s
=> CACHED [worker builder 9/9] RUN pip install --no-cache-dir --upgrade pip && pip install --no-cache-dir ti 0.0s
=> CACHED [worker final 5/9] COPY --from=builder /venv /venv 0.0s
=> CACHED [worker final 6/9] COPY --from=builder /models /app/models 0.0s
=> CACHED [worker final 7/9] COPY . /app/application 0.0s
=> CACHED [worker final 8/9] RUN mkdir -p /app/application/inputs/local 0.0s
=> CACHED [backend final 9/9] RUN chown -R appuser:appuser /app 0.0s
=> [worker] exporting to image 0.3s
=> => exporting layers 0.0s
=> => exporting manifest sha256:41b7093d7c6495fa315a26bd3dd38899e206f2c598537a8b330315314bc032b2 0.0s
=> => exporting config sha256:30dc4530fd3304e02b240334171db83926c4dbc7a61eef3cd81040470863a63f 0.0s
=> => exporting attestation manifest sha256:8642fd8c36272cbe5e35273e3ec200818982cd8aa99d6f6ef29dfbca359c12c4 0.1s
=> => exporting manifest list sha256:bd741fd1b3ed6ffb0be074b266f5a025c30e52181893f5538f44c270bbd39556 0.0s
=> => naming to docker.io/library/docsgpt-oss-worker:latest 0.0s
=> => unpacking to docker.io/library/docsgpt-oss-worker:latest 0.0s
=> [backend] exporting to image 0.3s
=> => exporting layers 0.0s
=> => exporting manifest sha256:bccc8b517ec8e4ba23efc3e1b71a16a109a80bdac942f014ee5f8ed439042d27 0.0s
=> => exporting config sha256:0169b06a2dfe81ce52971e7c5c2a7282d9ba2969482801ac344fdce75d7e7f60 0.0s
=> => exporting attestation manifest sha256:31ede6053a59ccc31f855dc1243bfacac709a536815cd98ac98ae64546d31f74 0.1s
=> => exporting manifest list sha256:f2b33ac9a260b5186f864e11159f942d6139b1c992ecffde6d42262a7e1f2e56 0.0s
=> => naming to docker.io/library/docsgpt-oss-backend:latest 0.0s
=> => unpacking to docker.io/library/docsgpt-oss-backend:latest 0.0s
=> [worker] resolving provenance for metadata file 0.0s
=> [backend] resolving provenance for metadata file 0.0s
=> [frontend internal] load build definition from Dockerfile 0.0s
=> => transferring dockerfile: 201B 0.0s
=> [frontend internal] load metadata for docker.io/library/node:20.6.1-bullseye-slim 0.9s
=> [frontend auth] library/node:pull token for registry-1.docker.io 0.0s
=> [frontend internal] load .dockerignore 0.0s
=> => transferring context: 2B 0.0s
=> [frontend 1/5] FROM docker.io/library/node:20.6.1-bullseye-slim@sha256:ee905d8492c443aebe41f4cc525ebabefef757 0.1s
=> => resolve docker.io/library/node:20.6.1-bullseye-slim@sha256:ee905d8492c443aebe41f4cc525ebabefef757df43556c4 0.1s
=> [frontend internal] load build context 0.2s
=> => transferring context: 2.46MB 0.2s
=> CACHED [frontend 2/5] WORKDIR /app 0.0s
=> CACHED [frontend 3/5] COPY package*.json ./ 0.0s
=> CACHED [frontend 4/5] RUN npm install 0.0s
=> CACHED [frontend 5/5] COPY . . 0.0s
=> [frontend] exporting to image 0.1s
=> => exporting layers 0.0s
=> => exporting manifest sha256:aa77fd68e595b54e232fe831640972c18c311e4763d4738c164cdcf82959ac9b 0.0s
=> => exporting config sha256:6ea5e684bb76995f8340c9a0fecc7bc8a26cb60d9faf95f1572fb351c75d8412 0.0s
=> => exporting attestation manifest sha256:66a043a6ab89c95dea956e195fe064d6bddffd6ffe1c417fe3bc522976f4de96 0.0s
=> => exporting manifest list sha256:765ed4b848381114866573838f0c96b91a52f9ccf69639c6a149ae73e30252b5 0.0s
=> => naming to docker.io/library/docsgpt-oss-frontend:latest 0.0s
=> => unpacking to docker.io/library/docsgpt-oss-frontend:latest 0.0s
=> [frontend] resolving provenance for metadata file 0.0s
[+] Running 8/8
✔ backend Built 0.0s
✔ frontend Built 0.0s
✔ worker Built 0.0s
✔ Container docsgpt-oss-redis-1 Started 1.3s
✔ Container docsgpt-oss-mongo-1 Started 1.4s
✔ Container docsgpt-oss-worker-1 Started 1.8s
✔ Container docsgpt-oss-backend-1 Started 1.8s
✔ Container docsgpt-oss-frontend-1 Started 1.7s
DocsGPT is now configured to use OpenAI on http://localhost:5173
You can stop the application by running: docker compose -f "K:\DocsGPT\deployment\docker-compose.yaml" down
DocsGPT Setup Complete.
👀 Have you spent some time to check if this bug has been raised before?
- [x] I checked and didn't find similar issue
🔗 Are you willing to submit PR?
No
🧑⚖️ Code of Conduct
- [x] I agree to follow this project's Code of Conduct
I think Ill try to add more details to this error, to make it easier to debug
I ran into same error on Windows 11 after fresh install @dartpain.
- Serve Local (with Ollama)
- serve Ollama with CPU
An other Issue I had to solve: "setup.ps1" creates an .env-File, that was UTF-8-BOM encoded (which leads to Error unexpected character "\ufeff" in variable name "\ufeffAPI_KEY=xxxx\r"), instead of UTF-8.
same problem, i asked ai and it said the 'chunks' parameter expected by the backend should be an number rather than a string, maybe it could fix the problem?
If anyone still has a similar issue, pease check your backend logs, as these errors may be happening for different reasons.
check the line in logs that starts with /stream - error
Also please make sure you are using latest version by running git pull
I had this error too when choosing ollama for my Inference Engine
started with this fix: https://docs.docsgpt.cloud/Models/local-inference The OPENAI_BASE_URL examples above use http://localhost. If you are running DocsGPT within Docker and your local inference engine is running on your host machine (outside of Docker), you will likely need to replace localhost with http://host.docker.internal to ensure Docker can correctly access your host's services. For example, http://host.docker.internal:11434/v1 for Ollama.
In the deployment/docker-compose.yaml file MODEL_NAME is missing from backend environment. Changed LLM_PROVIDER during troubleshooting since it did not exist in .env file
backend:
user: root
build: ../application
environment:
- API_KEY=$API_KEY
- EMBEDDINGS_KEY=$API_KEY
- LLM_PROVIDER=$LLM_NAME
- LLM_NAME=$LLM_NAME
- MODEL_NAME=$MODEL_NAME
Added MODEL_NAME to Settings class in application/core/settings.py
added line 18: MODEL_NAME: Optional[str] = None
Updated get_gpt_model() function in application/utils.py
modified line 31: return settings.MODEL_NAME or settings.LLM_NAME or model_map.get(settings.LLM_PROVIDER, "")
Hey @dartpain, if this issue remains unfixed, can you please assign this issue to me, would be happy to work on it
I have been using the /stream endpoint from quite a while and it seems to work fine. I am attaching some screenshots for the same.
The API also works fine, I think there might be some mismatch in some JSON Payload which is giving your error. Can anyone help me reciprocate that error?