langgenius/dify
1
0
Fork 0
mirror of https://github.com/langgenius/dify.git synced 2026-07-21 12:05:30 -04:00
Code Issues 775 Packages Projects Releases 10 Wiki Activity

The workflow does not support configuring the response results for streaming HTTP requests. #21015

New Issue
Closed
opened 2026-02-21 20:10:15 -05:00 by yindo · 2 comments
yindo commented 2026-02-21 20:10:15 -05:00
Owner
Copy Link
Copy Source

Originally created by @t00011666 on GitHub (Dec 12, 2025).

Self Checks

  • I have read the Contributing Guide and Language Policy.
  • This is only for bug report, if you would like to ask a question, please head to Discussions.
  • I have searched for existing issues search for existing issues, including closed ones.
  • I confirm that I am using English to submit this report, otherwise it will be closed.
  • 【中文用户 & Non English User】请使用英语提交，否则会被关闭 ：）
  • Please do not modify this template :) and fill in all the required fields.

Dify version

1.9.2

Cloud or Self Hosted

Self Hosted (Docker)

Steps to reproduce

  1. Create a workflow -> Configure the start node -> Configure an HTTP interface that supports streaming output -> Configure the end node.
  2. Ask a question.
  3. Output the result.

✔️ Expected Behavior

The workflow supports configuring the streaming output of HTTP request responses; the end node can display the response results in a streaming manner, creating an effect similar to typing text on screen.

❌ Actual Behavior

The workflow does not support configuring the response results for streaming HTTP requests; therefore, the end node will continuously wait for the HTTP request to complete before displaying the response results.

Originally created by @t00011666 on GitHub (Dec 12, 2025). ### Self Checks - [x] I have read the [Contributing Guide](https://github.com/langgenius/dify/blob/main/CONTRIBUTING.md) and [Language Policy](https://github.com/langgenius/dify/issues/1542). - [x] This is only for bug report, if you would like to ask a question, please head to [Discussions](https://github.com/langgenius/dify/discussions/categories/general). - [x] I have searched for existing issues [search for existing issues](https://github.com/langgenius/dify/issues), including closed ones. - [x] I confirm that I am using English to submit this report, otherwise it will be closed. - [x] 【中文用户 & Non English User】请使用英语提交，否则会被关闭 ：） - [x] Please do not modify this template :) and fill in all the required fields. ### Dify version 1.9.2 ### Cloud or Self Hosted Self Hosted (Docker) ### Steps to reproduce 1. Create a workflow -> Configure the start node -> Configure an HTTP interface that supports streaming output -> Configure the end node. 2. Ask a question. 3. Output the result. ### ✔️ Expected Behavior The workflow supports configuring the streaming output of HTTP request responses; the end node can display the response results in a streaming manner, creating an effect similar to typing text on screen. ### ❌ Actual Behavior The workflow does not support configuring the response results for streaming HTTP requests; therefore, the end node will continuously wait for the HTTP request to complete before displaying the response results.
yindo added the 💪 enhancement label 2026-02-21 20:10:15 -05:00
yindo closed this issue 2026-02-21 20:10:15 -05:00
yindo commented 2026-02-21 20:10:16 -05:00
Author
Owner
Copy Link
Copy Source

@dosubot[bot] commented on GitHub (Dec 12, 2025):

You're correct—Dify workflows (including v1.9.2, self-hosted) do not support streaming HTTP response results in the end node. Even if your HTTP node or external API streams data (like SSE), the workflow engine buffers the entire response and only delivers it to the end node after completion, so you won't see incremental updates or a typing effect in the UI. This is a known architectural limitation, confirmed by maintainers and multiple user reports (example, example, example).

The core reason is that the end node collects all outputs at once after streaming is complete, as shown in the codebase (EndNode _run method). While the workflow engine supports streaming chunk events at the event level (ResponseStreamCoordinator), the end node only materializes outputs after all upstream streaming is finished.

There are no configuration options or official workarounds to enable true streaming for HTTP response results in workflows. Some users have built custom proxy adapters (like OpenAI-Stream-HTTP-Adapter) to convert SSE streams into OpenAI-compatible streaming responses, which can be consumed by LLM nodes that support streaming, but this does not enable streaming for generic HTTP nodes.

If streaming from arbitrary APIs is critical for your use case, it would require significant custom development—such as building a custom node or plugin that mimics the LLM streaming pattern (discussion). There is significant user demand for this feature, but as of now, no official roadmap commitment to support it.

For updates or to track progress, you can follow these related issues and discussions:

To reply, just mention @dosu.

How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other

Chat with Dosu Open in Cursor Join Discord Share on X

@dosubot[bot] commented on GitHub (Dec 12, 2025): <!-- Answer --> You're correct—Dify workflows (including v1.9.2, self-hosted) do not support streaming HTTP response results in the end node. Even if your HTTP node or external API streams data (like SSE), the workflow engine buffers the entire response and only delivers it to the end node after completion, so you won't see incremental updates or a typing effect in the UI. This is a known architectural limitation, confirmed by maintainers and multiple user reports ([example](https://github.com/langgenius/dify/issues/28083), [example](https://github.com/langgenius/dify/issues/24578), [example](https://github.com/langgenius/dify/discussions/21711)). The core reason is that the end node collects all outputs at once after streaming is complete, as shown in the codebase ([EndNode _run method](https://github.com/langgenius/dify/blob/f4c7f98a01e6ca1a4271cb1cc63d45693ab94cfd/api/core/workflow/nodes/end/end_node.py#L16-L47)). While the workflow engine supports streaming chunk events at the event level ([ResponseStreamCoordinator](https://github.com/langgenius/dify/blob/f4c7f98a01e6ca1a4271cb1cc63d45693ab94cfd/api/core/workflow/graph_engine/response_coordinator/coordinator.py#L71-L602)), the end node only materializes outputs after all upstream streaming is finished. There are no configuration options or official workarounds to enable true streaming for HTTP response results in workflows. Some users have built custom proxy adapters (like [OpenAI-Stream-HTTP-Adapter](https://github.com/gitsrc/OpenAI-Stream-HTTP-Adapter)) to convert SSE streams into OpenAI-compatible streaming responses, which can be consumed by LLM nodes that support streaming, but this does not enable streaming for generic HTTP nodes. If streaming from arbitrary APIs is critical for your use case, it would require significant custom development—such as building a custom node or plugin that mimics the LLM streaming pattern ([discussion](https://github.com/langgenius/dify/discussions/21711)). There is significant user demand for this feature, but as of now, no official roadmap commitment to support it. For updates or to track progress, you can follow these related issues and discussions: - [Streaming output limitation in workflows](https://github.com/langgenius/dify/issues/28083) - [SSE/HTTP streaming support discussion](https://github.com/langgenius/dify/issues/24578) - [Streaming from non-LLM APIs](https://github.com/langgenius/dify/discussions/21711) <!-- Dosu Comment Footer --> *To reply, just mention [@dosu](https://go.dosu.dev/dosubot).* --- <sup>How did I do? [Good](https://app.dosu.dev/response-feedback/d0d2dfb0-6ee7-41d9-966d-e32741a018c0?feedback_type=great_response) | [Irrelevant](https://app.dosu.dev/response-feedback/d0d2dfb0-6ee7-41d9-966d-e32741a018c0?feedback_type=irrelevant_answer) | [Incorrect](https://app.dosu.dev/response-feedback/d0d2dfb0-6ee7-41d9-966d-e32741a018c0?feedback_type=incorrect_sources) | [Verbose](https://app.dosu.dev/response-feedback/d0d2dfb0-6ee7-41d9-966d-e32741a018c0?feedback_type=too_verbose) | [Hallucination](https://app.dosu.dev/response-feedback/d0d2dfb0-6ee7-41d9-966d-e32741a018c0?feedback_type=hallucination) | [Report 🐛](https://app.dosu.dev/response-feedback/d0d2dfb0-6ee7-41d9-966d-e32741a018c0?feedback_type=bug_report) | [Other](https://app.dosu.dev/response-feedback/d0d2dfb0-6ee7-41d9-966d-e32741a018c0?feedback_type=other)</sup> [![Chat with Dosu](https://dosu.dev/dosu-chat-badge.svg)](https://app.dosu.dev/6088a4f4-073a-43f3-b3d4-4f06fb0c05a0/ask?utm_source=github)&nbsp;[![Open in Cursor](https://dosu.dev/dosu-in-cursor.svg)](https://cursor.com/link/prompt?text=You%27re%20correct%E2%80%94Dify%20workflows%20%28including%20v1.9.2%2C%20self-hosted%29%20do%20not%20support%20streaming%20HTTP%20response%20results%20in%20the%20end%20node.%20Even%20if%20your%20HTTP%20node%20or%20external%20API%20streams%20data%20%28like%20SSE%29%2C%20the%20workflow%20engine%20buffers%20the%20entire%20response%20and%20only%20delivers%20it%20to%20the%20end%20node%20after%20completion%2C%20so%20you%20won%27t%20see%20incremental%20updates%20or%20a%20typing%20effect%20in%20the%20UI.%20This%20is%20a%20known%20architectural%20limitation%2C%20confirmed%20by%20maintainers%20and%20multiple%20user%20reports%20%28%5Bexample%5D%28https%3A//github.com/langgenius/dify/issues/28083%29%2C%20%5Bexample%5D%28https%3A//github.com/langgenius/dify/issues/24578%29%2C%20%5Bexample%5D%28https%3A//github.com/langgenius/dify/discussions/21711%29%29.%0A%0AThe%20core%20reason%20is%20that%20the%20end%20node%20collects%20all%20outputs%20at%20once%20after%20streaming%20is%20complete%2C%20as%20shown%20in%20the%20codebase%20%28%5BEndNode%20_run%20method%5D%28https%3A//github.com/langgenius/dify/blob/f4c7f98a01e6ca1a4271cb1cc63d45693ab94cfd/api/core/workflow/nodes/end/end_node.py%23L16-L47%29%29.%20While%20the%20workflow%20engine%20supports%20streaming%20chunk%20events%20at%20the%20event%20level%20%28%5BResponseStreamCoordinator%5D%28https%3A//github.com/langgenius/dify/blob/f4c7f98a01e6ca1a4271cb1cc63d45693ab94cfd/api/core/workflow/graph_engine/response_coordinator/coordinator.py%23L71-L602%29%29%2C%20the%20end%20node%20only%20materializes%20outputs%20after%20all%20upstream%20streaming%20is%20finished.%0A%0AThere%20are%20no%20configuration%20options%20or%20official%20workarounds%20to%20enable%20true%20streaming%20for%20HTTP%20response%20results%20in%20workflows.%20Some%20users%20have%20built%20custom%20proxy%20adapters%20%28like%20%5BOpenAI-Stream-HTTP-Adapter%5D%28https%3A//github.com/gitsrc/OpenAI-Stream-HTTP-Adapter%29%29%20to%20convert%20SSE%20streams%20into%20OpenAI-compatible%20streaming%20responses%2C%20which%20can%20be%20consumed%20by%20LLM%20nodes%20that%20support%20streaming%2C%20but%20this%20does%20not%20enable%20streaming%20for%20generic%20HTTP%20nodes.%0A%0AIf%20streaming%20from%20arbitrary%20APIs%20is%20critical%20for%20your%20use%20case%2C%20it%20would%20require%20significant%20custom%20development%E2%80%94such%20as%20building%20a%20custom%20node%20or%20plugin%20that%20mimics%20the%20LLM%20streaming%20pattern%20%28%5Bdiscussion%5D%28https%3A//github.com/langgenius/dify/discussions/21711%29%29.%20There%20is%20significant%20user%20demand%20for%20this%20feature%2C%20but%20as%20of%20now%2C%20no%20official%20roadmap%20commitment%20to%20support%20it.%0A%0AFor%20updates%20or%20to%20track%20progress%2C%20you%20can%20follow%20these%20related%20issues%20and%20discussions%3A%0A-%20%5BStreaming%20output%20limitation%20in%20workflows%5D%28https%3A//github.com/langgenius/dify/issues/28083%29%0A-%20%5BSSE/HTTP%20streaming%20support%20discussion%5D%28https%3A//github.com/langgenius/dify/issues/24578%29%0A-%20%5BStreaming%20from%20non-LLM%20APIs%5D%28https%3A//github.com/langgenius/dify/discussions/21711%29)&nbsp;[![Join Discord](https://img.shields.io/badge/join-5865F2?logo=discord&logoColor=white&label=)](https://go.dosu.dev/discord-bot)&nbsp;[![Share on X](https://img.shields.io/badge/X-share-black)](https://twitter.com/intent/tweet?text=%40dosu_ai%20helped%20me%20solve%20this%20issue!&url=https%3A//github.com/langgenius/dify/issues/29597)
yindo commented 2026-02-21 20:10:16 -05:00
Author
Owner
Copy Link
Copy Source

@crazywoola commented on GitHub (Dec 12, 2025):

Please see this https://github.com/langgenius/dify/issues/29277 as well.

The bot is correct about this. This is an expected behavior.

@crazywoola commented on GitHub (Dec 12, 2025): Please see this https://github.com/langgenius/dify/issues/29277 as well. The bot is correct about this. This is an expected behavior.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
Knowledge Pipeline
Proposal
WIP
cloud
dependencies
feat:webapp
good first issue
outdated
performance
pull-request
Mirrored from GitHub Pull Request
 python:uv
refactor
size:L
stale
status: accepting prs
status: blocked
status: in discussion
stt & tts
vulnerablity
web
wont-fix
🌊 feat:workflow
🌚 invalid
🐞 bug
👻 feat:rag
💪 enhancement
📚 documentation
🔨 feat:tools
🙋‍♂️ question
🙏 help wanted
🤔 cant-reproduce
🤖 feat:agent
No Label 💪 enhancement
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: langgenius/dify#21015
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?