## Summary
- Add custom `base_url` support to the Google Go model driver.
- Preserve Google URL suffix configuration when creating custom base URL
driver instances.
- Validate Google chat/stream request inputs before constructing the SDK
client.
- Cover Google model listing, connection checks, base URL resolution,
and request validation with focused tests.
## What changed
- `GoogleModel.NewInstance` now returns a Google driver configured with
the supplied base URL map.
- Google SDK client creation now resolves configured base URLs through
`genai.HTTPOptions.BaseURL`.
- Base URL lookup supports configured regions, empty-region keys, and
`default` fallback.
- Google chat, streaming chat, embeddings, and model listing now reject
blank API keys before creating SDK clients.
- Google chat and streaming chat now reject blank model names locally,
and streaming chat rejects a nil sender.
- Existing message handling, embeddings, pagination, and provider errors
are preserved.
## Why
Google custom model instances could not use configured base URLs because
`NewInstance` returned `nil` and the SDK client path ignored the driver
base URL map. The request validation keeps invalid Google calls from
reaching SDK client construction with blank credentials or incomplete
chat inputs.
### What problem does this PR solve?
The Go GPUStack driver returned a stub error for `Embed()` even though
GPUStack exposes OpenAI-compatible embeddings on the **v1-openai** route
(not `v1/embeddings`).
### Type of change
- [x] New Feature (non-breaking change which adds functionality)
## Summary
- Validate Hunyuan embedding model name and API key before building
requests.
- Reuse region-aware base URL validation for embedding requests.
- Replace the stale unsupported Embed test with happy-path and
validation coverage.
## What changed
- Added early Hunyuan Embed validation for missing model names and API
keys.
- Routed Embed through the same base URL region guard used by the other
Hunyuan methods.
- Updated Hunyuan tests to configure the embedding suffix and cover
Embed success plus invalid inputs.
## Why
Hunyuan Embed is implemented, but the existing test still expected it to
be unsupported and could panic before returning a normal validation
error. This keeps the implemented embedding path aligned with the
current driver behavior and prevents nil input panics.
Closes#15087
Refs #14736
### Summary
Closes#15381
Every provider in `internal/entity/models/` reads its streaming response
with `bufio.NewScanner(resp.Body)` and iterates over `scanner.Scan()`.
The default `bufio.Scanner` maximum token size is 64KB, so when an
upstream sends a single SSE `data:` line larger than 64KB (long content
deltas, large tool or function call argument blobs, bundled
`reasoning_content`, or providers that emit a whole message in one
event) `scanner.Scan()` returns `false` and `scanner.Err()` returns
`bufio.ErrTooLong`. Streaming chat then ends with an error partway
through the response.
This change adds `scanner.Buffer(make([]byte, 64*1024), 1024*1024)`
immediately after every SSE scanner that was still bare, raising the cap
to 1MB. 1MB is the value already used for streaming chat in `openai.go`,
`modelscope.go`, `groq.go`, `mistral.go`, `xai.go` and the other already
patched providers (the 8MB cap in the repo is reserved for TTS and
embedding paths), so this simply converges the remaining providers onto
the established pattern. Nothing else changes: line parsing, `data:`
prefix handling, `[DONE]` detection, JSON unmarshalling, error handling,
and the existing `scanner.Err()` checks all stay the same.
Providers covered (23 scanners across 22 files): 302ai, aliyun,
baichuan, baidu, cohere, deepinfra, deepseek, gitee, huggingface,
lmstudio, minimax (the chat scanner, whose TTS scanner was already
bumped), moonshot, nvidia, ollama, openrouter, orcarouter, paddleocr,
siliconflow, tokenhub, vllm, volcengine, xunfei, zhipu-ai. `jiekouai.go`
is excluded because it is covered by the in flight #15337.
A table driven regression test (`sse_scanner_buffer_test.go`) streams a
single 128KB `data:` content delta followed by `data: [DONE]` through an
`httptest` server and asserts that `ChatStreamlyWithSender` delivers the
full content with no error across a representative subset of providers.
Without the buffer fix the test fails with `bufio.Scanner: token too
long`.
This PR also removes three duplicate declarations of the package level
`roundTripperFunc` test helper that several recently merged provider PRs
each added independently, which had left the `internal/entity/models`
test package unable to compile. The helper now lives in a single place
and is shared.
### Type of change
- [x] Bug Fix (non-breaking change which fixes an issue)
### What problem does this PR solve?
Added 4 new models:
deepseek-ai/DeepSeek-V4-Pro
deepseek-ai/DeepSeek-V4-Flash
Pro/moonshotai/Kimi-K2.6
Pro/zai-org/GLM-5.1
### Type of change
- [x] New Feature (non-breaking change which adds functionality)
## Summary
- Harden `NewNovitaModel` to avoid panics when `http.DefaultTransport`
is a custom non-`*http.Transport` RoundTripper.
- Fallback to a safe transport (`ProxyFromEnvironment`) while preserving
existing pooling/timeout settings.
Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
## Summary
- Harden JieKouAI request validation before outbound provider calls
- Force non-streaming and streaming chat methods to use their expected
stream modes
- Make model listing use a bodyless GET and parse model responses
without panics
Closes#14736
---------
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
## Summary
- Harden `NewModelScopeModel` to avoid panics when
`http.DefaultTransport` is a custom non-`*http.Transport` RoundTripper.
- Fallback to a safe transport (`ProxyFromEnvironment`) while preserving
existing pooling/timeout settings.
- Add `TestModelScopeNewModelWithCustomDefaultTransport` regression
coverage.
Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
## Summary
- Harden `NewVoyageModel` to avoid panics when `http.DefaultTransport`
is a custom non-`*http.Transport` RoundTripper.
- Fallback to a safe transport (`ProxyFromEnvironment`) while preserving
existing pooling/timeout settings.
- Add `TestVoyageNewModelWithCustomDefaultTransport` regression
coverage.
Co-authored-by: Cursor <cursoragent@cursor.com>
## Summary
- Harden `NewLongCatModel` to avoid panics when `http.DefaultTransport`
is a custom non-`*http.Transport` RoundTripper.
- Fallback to a safe transport (`ProxyFromEnvironment`) while preserving
existing pooling/timeout settings.
- Add `TestLongCatNewModelWithCustomDefaultTransport` regression
coverage.
Co-authored-by: Cursor <cursoragent@cursor.com>
### What problem does this PR solve?
implement delete, rebuild api for connector
### Type of change
- [x] New Feature (non-breaking change which adds functionality)
## Summary
- Harden the 302.AI model driver request validation and response parsing
paths.
- Add focused tests for chat request mode, model listing, malformed
provider responses, and input validation.
## What changed
- Validate API keys, model names, rerank queries, ASR file paths, OCR
inputs, parse URLs, task IDs, and model-list IDs before use.
- Keep chat and streaming methods from accepting conflicting `stream`
values in request payloads.
- Send `ListModels` as a bodyless GET and parse the response with typed
JSON structs instead of unchecked assertions.
- Remove raw SSE event logging from stream handling.
## Why
The driver could panic or send inconsistent requests when optional
config fields were nil, empty, malformed, or contradicted the method
path. This keeps provider-driver behavior explicit while preserving the
existing supported 302.AI flows.
Closes#14736
### What problem does this PR solve?
implement create_connector API
### Type of change
- [x] New Feature (non-breaking change which adds functionality)
## Summary
- centralize TokenHub chat request validation for chat and streaming
calls
- reject blank TokenHub model names before sending provider requests
- send TokenHub model listing requests as bodyless GET requests
## What changed
- Added shared TokenHub chat request validation for API key, model name,
and messages.
- Updated `ListModels` to call `GET /models` without a request body.
- Added focused tests for blank model names and accidental GET request
bodies.
- Replaced an httptest handler callback `t.Fatalf` with `t.Errorf` plus
an HTTP error and return.
## Why
TokenHub chat requests should fail locally for invalid model names
instead of sending avoidable malformed requests upstream. Model listing
should also match normal GET semantics and avoid sending an empty JSON
body.
Closes#14736
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
### What problem does this PR solve?
As title
### Type of change
- [x] Bug Fix (non-breaking change which fixes an issue)
- [x] New Feature (non-breaking change which adds functionality)
- [x] Refactoring
### What problem does this PR solve?
`ReplicateModel.Rerank` in `internal/entity/models/replicate.go` was a
`"replicate, no such method"` stub. The chat path landed in #14958 and
the embed path in #15073; rerank is the last major retrieval surface
still missing on this provider.
Until this PR, a tenant who selected a Replicate reranker model got the
sentinel error on every rerank call.
Co-authored-by: sxxtony <sxxtony@users.noreply.github.com>
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
### What problem does this PR solve?
This PR adds Qiniu provider integration for the Go model driver layer in
RAGFlow.
Supported capabilities:
- [X] Chat
- [X] Think Chat
- [X] Stream Chat
- [X] Stream Think Chat
- [X] Model listing
- [X] Provider configuration and factory registration
Verified examples from the CLI:
```
login user '***' password '***';
ADD PROVIDER 'qiniu';
CREATE PROVIDER 'qiniu' INSTANCE 'test' KEY '***';
chat with 'deepseek/deepseek-v3.1-terminus-thinking@test@qiniu' message
'hello';
think chat with 'deepseek/deepseek-v3.1-terminus-thinking@test@qiniu'
message 'hello';
stream chat with 'deepseek/deepseek-v3.1-terminus-thinking@test@qiniu'
message 'hello, what are you';
stream think chat with
'deepseek/deepseek-v3.1-terminus-thinking@test@qiniu' message 'hello,
what are you';
stream think chat with 'qwen3-max-2026-01-23@test@qiniu' message 'hello,
what are you';
LIST MODELS FROM 'qiniu' 'test';
```
### Type of change
- [X] New Feature
- [X] Provider integration
## Summary
- add the VolcEngine `models` URL suffix used by the existing Go
`ListModels` implementation
- return a clear error when the VolcEngine models suffix is missing
- add focused VolcEngine model-listing regression tests
## What changed
- Added `url_suffix.models` to `conf/models/volcengine.json`.
- Normalized the configured models suffix before building the request
URL.
- Covered config loading, successful model listing, upstream errors, and
missing suffix handling.
## Why
`VolcEngine.ListModels` already builds requests from `URLSuffix.Models`,
but the bundled VolcEngine config did not define that suffix. That left
the model-listing path unable to call the documented `/models` endpoint
from the existing provider config.
Fixes#14701
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
## Summary
- route hosted MinerU.Net and PaddleOCR.Net provider names to their
existing Go drivers
- add regression coverage for loading the hosted OCR provider configs
through ProviderManager
## What changed
- Added canonical provider-name aliases for the hosted OCR provider
display names.
- Covered both bundled configs with a focused provider-manager test.
## Why
The hosted provider configs use display names with `.Net`, while model
factory dispatch lowercases the provider name. Without aliases, those
configs fall through to `DummyModel` instead of using the existing
MinerU and PaddleOCR drivers.
---------
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
## Summary
- Add LongCat model-list support through the documented
OpenAI-compatible models endpoint.
## What changed
- Add the LongCat `models` URL suffix for `/openai/v1/models`.
- Implement `ListModels` for the LongCat Go driver.
- Delegate `CheckConnection` to the lightweight model-list request.
- Add focused regression coverage for successful, malformed, oversized,
and missing-key responses.
## Why
LongCat documents a models endpoint under the OpenAI-compatible API
surface, but the Go driver still returned `no such method` for model
listing and connection checks.
## Validation
- `go test ./internal/entity/models -run TestLongCat -count=1`
- `go test -race ./internal/entity/models -run TestLongCat -count=1`
- `go test ./internal/entity -count=1`
- `git diff --check`
## Notes
- Related to the broader Go model provider tracking in #14736, but this
PR only handles LongCat model listing.
- `go test ./internal/entity/models -count=1` is currently blocked by an
unrelated Astraflow test panic outside this LongCat change.
---------
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
## Summary
- add the xAI `models` URL suffix used by the existing Go `ListModels`
implementation
- return a clear error when the xAI models suffix is missing
- add focused xAI model-listing and connection-check regression tests
## What changed
- Added `url_suffix.models` to `conf/models/xai.json`.
- Normalized the configured models suffix before building the request
URL.
- Covered config loading, successful model listing, upstream errors,
API-key validation, missing suffix handling, and `CheckConnection`
delegation.
## Why
`XAIModel.ListModels` already builds requests from `URLSuffix.Models`,
and `CheckConnection` delegates to that method. The bundled xAI config
did not define that suffix, which left the model-listing path unable to
call the provider `/models` endpoint from the existing provider config.
## Validation
- `go test ./internal/entity/models -run TestXAI -count=1`
- `go test ./internal/entity -count=1`
- `git diff HEAD~1..HEAD --check`
## Notes
- `go test ./internal/entity/models -count=1` currently fails in
unchanged Astraflow coverage: `TestAstraflowEmbedReturnsNoSuchMethod`
panics before reaching any xAI assertions.
---------
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
### What problem does this PR solve?
implement provider `OrcaRouter`
**The following functionalities are now supported:**
**Cohere:**
- [x] Chat / Think Chat / Stream Chat / Stream Think Chat
- [x] Model listing
- [x] TTS
- [ ] Balance
### Type of change
- [x] New Feature (non-breaking change which adds functionality)
---------
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
Closes#15040.
ModelScope was listed unchecked in the Go-rewrite tracker #14736 and
already had an llm_factories.json entry (tags: LLM) but no Go driver, so
the new Go API server could not route ModelScope instances. The Python
side has supported it through the OpenAI-compatible base at
rag/llm/chat_model.py:618 (ModelScopeChat), which requires a
user-supplied base URL and appends /v1.
This adds:
- internal/entity/models/modelscope.go: self-hosted OpenAI-compatible
driver with chat (sync + SSE stream with idle-timeout cancellation),
list_models, and check_connection. Auth header is optional, matching the
xinference pattern, so deployments without auth and auth-enabled
deployments both work. Base URL is normalized so users can configure
either the root endpoint or the /v1 endpoint.
- internal/entity/models/modelscope_test.go: 12 tests covering name, URL
normalization, factory routing, chat happy path / auth header /
reasoning_content extraction, stream happy path / stream=false rejection
/ idle cancellation, list_models + check_connection, missing-base-URL
clear error, and the no-such-method sentinels.
- conf/models/modelscope.json: shipped config (class: "local",
url_suffix v1/chat/completions and v1/models).
- internal/entity/models/factory.go: case "modelscope" →
ModelScopeModel.
- internal/service/llm.go: ModelScope added to the selfDeployed map
alongside Ollama, Xinference, LocalAI, LM-Studio, GPUStack — the Python
side requires user-supplied URL with no default, so the Go side
classifies it the same way.
Follow-on issues will add Embed and Rerank, in line with how Novita,
NVIDIA, TogetherAI, and other providers landed method-by-method.
---------
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
### What problem does this PR solve?
This PR adds HuaweiCloud provider integration in RAGFlow.
Supported capabilities:
- [x] Chat / Think Chat / Stream Chat / Stream Think Chat
- [x] Embedding
- [x] Rerank
- [x] Model listing
- [x] Provider connection checking
Verified examples from the CLI:
```
check instance 'test' from 'HuaweiCloud';
chat with 'deepseek-v4-flash@test@HuaweiCloud' message 'hello';
think chat with 'deepseek-v4-flash@test@HuaweiCloud' message 'hello';
stream chat with 'deepseek-v4-flash@test@HuaweiCloud' message 'hello';
stream think chat with 'deepseek-v4-flash@test@HuaweiCloud' message
'hello';
embed text 'what is rag' 'who are you' with 'bge-m3@test@HuaweiCloud'
dimension 1024;
rerank query 'what is rag' document 'rag is retrieval augmented
generation' 'rag need llm' 'famous rag
project includes ragflow' with 'bge-reranker-v2-m3@test@HuaweiCloud' top
3;
list supported models from 'HuaweiCloud' 'test';
LIST MODELS FROM 'HuaweiCloud' 'test';
```
### Type of change
- [x] New Feature
- [x] Provider integration
## Summary
- Wire the Go TokenHub provider through the model factory.
- Harden TokenHub request handling for chat, streaming, embeddings, and
model listing.
- Add focused TokenHub unit coverage for factory wiring and provider
behavior.
## Notes
- Refs #14736.
- Follows up #15159.
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
## Summary
Closes#15165.
Implements the AWS Bedrock model provider for the Go API server, tracked
under #14736. Adds Converse + Converse-Stream chat and foundation-model
listing, with SigV4 signing over a hand-rolled `net/http` path that
matches the established pattern in `internal/entity/models/` (no new
direct `go.mod` deps).
## Linked tracker
Tracked under #14736 (Implement model providers of RAGFlow API server in
Go). Closes#15165.
### What problem does this PR solve?
The Go DeepInfra driver returned a stub error for `Rerank()` even though
DeepInfra serves reranker models at `POST /v1/inference/{model}` with
`query`, `documents`, and a `scores[]` response.
### Type of change
- [x] New Feature (non-breaking change which adds functionality)
Co-authored-by: Cursor <cursoragent@cursor.com>
### What problem does this PR solve?
Add a Go driver for **FuturMix** (https://futurmix.ai/docs), one of the
unchecked providers on the umbrella tracking issue #14736. FuturMix is
documented as an "OpenAI-compatible API" aggregator over Claude / GPT /
Gemini / DeepSeek (~22 models per their `/models` page).
Until this PR, a tenant who configured `futurmix` as a model provider in
the Go layer fell through to the default branch of
`internal/entity/models/factory.go` and got the dummy driver.
---------
Co-authored-by: sxxtony <sxxtony@users.noreply.github.com>
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
### What problem does this PR solve?
Closes#15167.
The Baidu Go provider advertises OCR support through
`paddleocr-vl-0.9b`, but `BaiduModel.OCRFile` dereferenced required
inputs before validating them. Calling OCR with a missing API config,
API key, or model name could panic instead of returning a normal error.
This PR adds explicit input validation for those required values.
### Type of change
- [x] Bug Fix (non-breaking change which fixes an issue)
---------
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
### What problem does this PR solve?
Closes#15142.
ZhipuAI lists `glm-ocr` as an OCR model, but the Go driver still
returned `no such method` from `OCRFile`. This wires the advertised
model to Z.AI's documented `layout_parsing` endpoint and returns the
`md_results` Markdown output through the existing `OCRFileResponse.Text`
field.
This PR also adds focused tests for URL input, raw file-content base64
input, and validation errors.
### Type of change
- [x] Bug Fix (non-breaking change which fixes an issue)
- [ ] New Feature (non-breaking change which adds functionality)
- [ ] Documentation Update
- [ ] Refactoring
- [ ] Performance Improvement
- [ ] Other (please describe):
### Test
- [x] `go test -vet=off ./internal/entity/models -run
'TestZhipuAIOCRFile'`
### What problem does this PR solve?
IDK how to implement **`Ollama`** on #14580 but it's totally wrong.
This is the rewrite version for **`Ollama`**
**Verified from CLI**
```
# Embed
RAGFlow(user)> embed text 'what is rag' 'who are you' with 'nomic-embed-text:latest@test12@ollama' dimension 1024;
+-----------+-------+
| dimension | index |
+-----------+-------+
| 768 | 0 |
| 768 | 1 |
+-----------+-------+
# Chat
RAGFlow(user)> think chat with 'qwen3:0.6b@test12@ollama' message 'who r u'
Thinking: Okay, the user asked, "Who r u?" I need to respond appropriately. First, I should acknowledge their question. Since I'm an AI, I don't have a physical form, but I can confirm that I'm a large language model. I should keep the response friendly and offer help. Let me make sure I'm not making up any information and that the response is natural. Also, I should check for any typos and ensure clarity. Alright, that should cover it.
Answer: I'm an AI language model, and I don't have a physical form. However, I can tell you that I'm designed to assist with questions and tasks. How can I help you today?
Time: 2.914285
RAGFlow(user)> stream think chat with 'qwen3:0.6b@test12@ollama' message 'who r u'
Thinking: , the user asked, "Who are you?" I need to respond appropriately. Since I'm an AI assistant, I should mention that I don't have a physical form or a mind. I should also clarify that I can help with various tasks like answering questions or providing information. It's important to keep the response friendly and informative while maintaining the correct tone.
Answer: don't have a physical form or a mind, but I'm here to help with your questions or tasks! What can I do for you today?
Time: 1.740047
# LisyModels
RAGFlow(user)> list supported models from 'ollama' 'test12'
+-------------------------+
| model_name |
+-------------------------+
| nomic-embed-text:latest |
| qwen3:0.6b |
+-------------------------+
```
### Type of change
- [x] Bug Fix (non-breaking change which fixes an issue)
- [x] Refactoring
### What problem does this PR solve?
Implement embed for Tencent Hunyuan
**Verified from CLI**
```
RAGFlow(user)> embed text 'what is rag' 'who are you' with 'hunyuan-embedding@test1@hunyuan' dimension 16;
+-----------+-------+
| dimension | index |
+-----------+-------+
| 1024 | 0 |
| 1024 | 1 |
+-----------+-------+
```
### Type of change
- [x] New Feature (non-breaking change which adds functionality)
- [x] Refactoring
### What problem does this PR solve?
Go: implement provider: PaddleOCR_Local
**Verified from CLI**
```
RAGFlow(user)> ocr with 'PaddleOCR-VL@test@paddleocr_local' file './internal/test1.jpg'
+----------------------+
| text |
+----------------------+
| ## Parallel to these |
+----------------------+
```
### Type of change
- [X] Bug Fix (non-breaking change which fixes an issue)
- [X] New Feature (non-breaking change which adds functionality)
- [X] Refactoring
## Summary
- Adds a `Hunyuan` Go driver so the new API server can route Tencent
Hunyuan chat instances (registered in `conf/llm_factories.json:3830` as
`Tencent Hunyuan`). Follows the same SaaS-driver shape used for
Astraflow, Avian, Novita, TogetherAI, Replicate, DeepInfra, Upstage, and
LongCat.
Closes#15087
---------
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
### What problem does this PR solve?
Closes#15102.
OpenAI's Go provider config advertises `whisper-1` as ASR and `tts-1` as
TTS, but the Go driver returned `openai, no such method` for both audio
paths and did not define `url_suffix.asr` / `url_suffix.tts`.
This PR:
- adds OpenAI audio URL suffixes for `audio/transcriptions` and
`audio/speech`
- implements non-streaming `TranscribeAudio` using multipart form
uploads
- implements non-streaming `AudioSpeech` using the OpenAI speech JSON
request shape
- keeps streaming TTS explicitly unsupported instead of sending binary
audio through the text SSE sender
- adds focused tests for config coverage, ASR/TTS request shape,
required TTS voice validation, and unsupported streaming TTS
### Type of change
- [x] Bug Fix (non-breaking change which fixes an issue)
---------
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
### What problem does this PR solve?
Go: implement reasoning_chat, TTS, ASR for Groq
**Verify from CLI**
```
RAGFlow(user)> think chat with 'qwen/qwen3-32b@test@groq' message 'who r u'
Thinking: Okay, the user asked, who r u. I need to determine what the user is asking. They may be asking about my identity. I should introduce my name and basic functions. The user might want to know what I can do, so I should list some common use cases, such as answering questions, creating writing, coding, and expressing opinions. The user may be curious about how they can interact with me, so they can be advised to ask any questions or provide instructions. Keep your answers conversational, avoid overly technical terms, keep answers concise, and encourage further interaction. Check if there's any ambiguity in the answer and make sure it's accurate and meets the user's needs. Also consider if there are other aspects the user may be interested in, such as my training data or performance. But since the question is basic, I'll focus on the essentials first and invite the user to ask more. In summary, respond to the user's questions by introducing yourself, your functions, and encouraging further interaction.
Answer: Hello! I'm Qwen. I am a large-scale language model developed by Tongyi Lab, designed to assist you in various ways, such as answering questions, creating text, logical reasoning, programming, and more. I aim to provide clear, accurate, and helpful information and support. How can I assist you today? Feel free to ask any questions or give me tasks! 😊
Time: 2.199908
RAGFlow(user)> stream think chat with 'openai/gpt-oss-20b@test@groq' message 'who r u'
Thinking: to respond politely.
Answer: ’m ChatGPT—an AI language model created by OpenAI. I’m here to answer questions, offer explanations, and help with a wide range of topics. How can I assist you today?
RAGFlow(user)> tts with 'canopylabs/orpheus-arabic-saudi@test@groq' text 'hello? show yourself' play format 'wav' param '{"voice": "fahad"}'
SUCCESS
RAGFlow(user)> asr with 'whisper-large-v3-turbo@test@groq' audio './internal/test.wav' param '{"language": "en"}'
+----------------------------------------------------------------------------------------------------------------------+
| text |
+----------------------------------------------------------------------------------------------------------------------+
| The examination and testimony of the experts enabled the Commission to conclude that five shots may have been fired |
+----------------------------------------------------------------------------------------------------------------------+
```
### Type of change
- [x] New Feature (non-breaking change which adds functionality)
### What problem does this PR solve?
Closes#15088.
Adds Groq support to the Go model-provider layer so Groq instances can
be routed through the Go API server with the same OpenAI-compatible
chat, streaming, model listing, and connection-check flow used by other
SaaS providers.
### Type of change
- [x] New Feature (non-breaking change which adds functionality)
## Summary
- Added a Groq Go model driver.
- Added the Groq provider catalog and default OpenAI-compatible API URL.
- Registered Groq in the model factory.
- Added focused provider tests.
## What changed
- Implemented chat completions, SSE streaming, ListModels, and
CheckConnection for Groq.
- Covered request shape, stream termination, reasoning fallback, model
listing, custom base URLs, safe transport setup, and unsupported
methods.
- Kept the provider catalog scoped to current Groq chat-capable model
IDs.
- Cleaned up pre-existing Go model package validation blockers so the
package can be tested normally with vet enabled.
## Why
The existing Python/provider catalog path includes Groq, but the Go
model-provider layer did not have a Groq driver, so the Go API server
could not instantiate or use Groq as requested in #15088.
## Notes
The model package now validates without disabling vet.
---------
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
## Summary
- Adds a `TokenPony` Go driver so the new API server can route TokenPony
chat instances, matching the existing Python `TokenPonyChat`
(`rag/llm/chat_model.py:1210`). Follows the same SaaS-driver shape used
for Astraflow, Avian, Novita, TogetherAI, Replicate, DeepInfra, Upstage,
and LongCat.
Closes#15086
---------
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
### What problem does this PR solve?
Fixes#15066
OpenRouter now exposes an official speech-to-text endpoint at `POST
/api/v1/audio/transcriptions`, but the Go model driver still returned
`openrouter, no such method` from `TranscribeAudio`. This left
OpenRouter ASR models unavailable through the Go API server even though
the provider already has OpenRouter audio support for TTS.
Related provider-tracking context: #14736
### Type of change
- [x] New Feature (non-breaking change which adds functionality)
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
### What problem does this PR solve?
This PR implements ASR and TTS support for the ZhipuAI Go driver.
The ZhipuAI model config already advertises `glm-asr-2512` as an ASR
model, but the Go driver returned `zhipu, no such method` from
`TranscribeAudio`. This adds the documented audio transcription endpoint
suffix and sends multipart transcription requests with `model`,
`stream=false`, and `file` fields.
Per maintainer review, this also adds the ZhipuAI TTS endpoint suffix
and implements `AudioSpeech` / `AudioSpeechWithSender` for `glm-tts`.
Closes#15133
### Type of change
- [x] Bug Fix (non-breaking change which fixes an issue)
- [x] New Feature (non-breaking change which adds functionality)
### What problem does this PR solve?
Closes#15089.
Adds PPIO support to the Go model-provider layer so PPIO instances can
be routed through the Go API server with the same OpenAI-compatible
chat, streaming, model listing, and connection-check flow used by other
SaaS providers.
### Type of change
- [x] New Feature (non-breaking change which adds functionality)
## Summary
- Added a PPIO Go model driver.
- Added the PPIO provider catalog and default OpenAI-compatible API URL.
- Registered PPIO in the model factory.
- Added focused provider and provider-manager tests.
## What changed
- Implemented chat completions, SSE streaming, ListModels, and
CheckConnection for PPIO.
- Covered request shape, stream termination, reasoning fallback, model
listing, custom base URLs, safe transport setup, unsupported methods,
and provider config loading.
- Kept the provider catalog aligned with the existing RAGFlow PPIO
factory model set.
- Cleaned up pre-existing Go model package validation blockers so the
scoped provider tests can run normally with vet enabled.
## Why
The existing Python/provider catalog path includes PPIO, but the Go
model-provider layer did not have a PPIO driver, so the Go API server
could not instantiate or use PPIO as requested in #15089.
### What problem does this PR solve?
implement rerank, asr, tts for TogetherAI
### Type of change
- [x] New Feature (non-breaking change which adds functionality)
### What problem does this PR solve?
implement ASR and TTS for Xinference
### Type of change
- [x] New Feature (non-breaking change which adds functionality)
- [x] Refactoring
### What problem does this PR solve?
`UpstageModel.ChatStreamlyWithSender` (in the driver merged via #14819)
only extracted `delta.content` from each SSE event. For the `solar-pro3`
reasoning family (and any future Upstage model that follows the same
wire shape), the chain-of-thought is streamed in a **separate
`delta.reasoning` field**, and the driver was silently dropping all of
it.
The non-streaming path already extracts `message.reasoning` into
`ChatResponse.ReasonContent` (added earlier in this PR's history), so
the same model produced **inconsistent behavior** between streaming and
non-streaming: a tenant calling `solar-pro3` with `reasoning_effort:
high` would see the reasoning trace if they used `ChatWithMessages` but
not if they used `ChatStreamlyWithSender`.
### Live evidence
Probed against `api.upstage.ai/v1/chat/completions` with `solar-pro3` +
`reasoning_effort: high` + `stream: true` (8000-token budget so the
reasoning has room to finish):
```
$ curl -sN -H "Authorization: Bearer <key>" -H "Content-Type: application/json" \
-X POST https://api.upstage.ai/v1/chat/completions \
-d '{"model":"solar-pro3","messages":[{"role":"user","content":"Compute 15% of 80."}],
"max_tokens":8000,"stream":true,"reasoning_effort":"high"}'
# across 168 SSE events:
# delta keys seen: [content reasoning role]
# delta.content total len: 121 chars (the visible answer)
# delta.reasoning total len: 159 chars (the chain-of-thought) <- driver dropped this
```
A representative event showing both fields side by side:
```json
data: {"choices":[{"index":0,"delta":{"reasoning":"15% = 0.15."}}]}
data: {"choices":[{"index":0,"delta":{"content":"15% of 80 is "}}]}
```
The 159 chars of reasoning were arriving on the wire and being thrown
away. `solar-pro2` was also probed (625 events); it does **not** emit
`delta.reasoning` — its reasoning is inlined into `delta.content` — so
this change is a no-op for it and for `solar-mini`.
### What this PR includes
- `internal/entity/models/upstage.go`: in the SSE scanner loop, extract
`delta.reasoning` before `delta.content` and forward each non-empty
chunk via the sender's second arg (the existing `reasonContent` channel
the non-stream path already populates).
The ordering contract is documented inline: reasoning chunks within a
single SSE event are emitted before content chunks, so a UI that pipes
both sees the chain-of-thought start before the answer for that token,
matching the wire order Upstage emits.
- `internal/entity/models/upstage_test.go`: three new tests pinning the
new behavior:
- `TestUpstageStreamExtractsReasoningDelta` — reasoning + content
forwarded to the right sender args; one-of invariant per call
- `TestUpstageStreamReasoningChunksArriveBeforeContent` — ordering
pinned within a single SSE event that carries both fields
- `TestUpstageStreamWithoutReasoningStillWorks` — regression net:
non-reasoning models (`solar-mini`, `solar-pro2`) continue to work; the
reason callback never fires
No interface change. No factory change. No config change.
### How was this tested?
```
$ go test -vet=off -run TestUpstage -count=1 -v ./internal/entity/models/...
... (existing tests 1..9 still pass) ...
=== RUN TestUpstageStreamExtractsReasoningDelta
--- PASS: TestUpstageStreamExtractsReasoningDelta (0.01s)
=== RUN TestUpstageStreamReasoningChunksArriveBeforeContent
--- PASS: TestUpstageStreamReasoningChunksArriveBeforeContent (0.01s)
=== RUN TestUpstageStreamWithoutReasoningStillWorks
--- PASS: TestUpstageStreamWithoutReasoningStillWorks (0.00s)
PASS
ok ragflow/internal/entity/models 0.034s
```
12/12 Upstage tests pass on go 1.25. `go build
./internal/entity/models/...` exits 0.
**Live integration test** (smoke test not committed) — the patched
driver was run directly against `api.upstage.ai/v1` with the same prompt
that produced the curl evidence above:
```
=== RUN TestUpstageStreamReasoningLiveSmoke
[OK] visible content: 50 chunks, 84 chars
[OK] reasoning: 39 chunks, 90 chars
content head 200: "\\(15\\% = \\frac{15}{100}=0.15\\).\n\n\\[\n0.15 \\times 80 = 12.\n\\]\n\n**15 % of 80 is 12.**"
reasoning head 200: "We need to compute 15% of 80. That's 0.15 * 80 = 12. So answer is 12. Provide explanation."
UPSTAGE STREAM REASONING SMOKE PASSED
--- PASS: TestUpstageStreamReasoningLiveSmoke (1.97s)
```
Before this fix, the same call would have produced **0 reasoning
chunks**. The 90 chars of reasoning that the patched driver now surfaces
are the chain-of-thought solar-pro3 emits when reasoning_effort is high.
### Type of change
- [x] Bug Fix (non-breaking change which fixes an issue)
### What problem does this PR solve?
`MistralModel.ChatWithMessages` (in the driver merged via #14807)
assumes that `choices[0].message.content` from `/v1/chat/completions` is
always a string and falls through to `return nil, fmt.Errorf("invalid
content format")` on anything else.
That assumption breaks for the **magistral reasoning family**
(`magistral-small-*`, `magistral-medium-*`). When the model needs a
chain-of-thought to answer, Mistral returns `content` as a **structured
array of typed parts**:
```json
"content": [
{"type": "thinking",
"thinking": [{"type": "text", "text": "Combined speed is 150 mph. 300 / 150 = 2 hours."}],
"closed": true},
{"type": "text", "text": "They will meet after **2 hours**."}
]
```
Concretely, this is what the live API returns today (probed against
`api.mistral.ai/v1`):
```
$ curl -H "Authorization: Bearer <key>" -H "Content-Type: application/json" \
-X POST https://api.mistral.ai/v1/chat/completions \
-d '{"model":"magistral-medium-latest",
"messages":[{"role":"user","content":"two trains 60mph and 90mph, 300mi apart, when do they meet? step by step."}],
"max_tokens":1024}'
HTTP 200
{ "choices":[{"message":{
"role":"assistant",
"content":[
{"type":"thinking","thinking":[{"type":"text","text":"Okay, let's see..."}],"closed":true},
{"type":"text","text":"To determine when the two trains meet..."}
]}}] }
```
With the current driver, every call like that returns the generic
`"invalid content format"` error. Trivial prompts that happen to fit in
a string answer still succeed, so the breakage is **non-deterministic
from the tenant's POV**: same model, same provider, sometimes works,
sometimes 500s with no useful error.
A secondary issue: `conf/models/mistral.json` does not include any
magistral model. The picker hid the broken path, which is why this
wasn't caught during #14807's review.
### What this PR includes
- New helper `extractMistralContent(raw interface{}) (answer,
reasonContent string, err error)` in
`internal/entity/models/mistral.go`, which normalizes both shapes
Mistral can return:
- `string` → historical path. `Answer = content`, `ReasonContent = ""`.
Preserves behavior for every non-reasoning model (`mistral-large-*`,
`mistral-small-*`, `ministral-*`, `codestral-*`, `pixtral-*`,
`open-mistral-nemo`).
- `[]interface{}` → walk the parts. Concatenate every `{"type":"text",
"text":...}` part into `Answer`; concatenate the inner text inside every
`{"type":"thinking", "thinking":[...]}` part into `ReasonContent`.
- `ChatWithMessages` now calls the helper instead of doing the raw
`.(string)` cast.
- Unknown part types are **skipped, not failed**. Mistral has been
adding new content variants quickly (audio chunks, citations, etc.);
this driver should not 500 every call when a new part type appears.
- `conf/models/mistral.json`: add `magistral-medium-latest` and
`magistral-small-latest`. Both are visible in `/v1/models` today.
No interface change. No factory change. No new dependencies.
### How was this tested?
**Unit tests** — 5 new tests in `internal/entity/models/mistral_test.go`
on top of the 27 already shipped via #14807:
- `TestMistralChatHandlesStringContent` — regression net for the
historical path
- `TestMistralChatExtractsReasoningFromStructuredContent` — the fixture
body is a trimmed copy of the actual `magistral-medium-latest` response
captured above; asserts both `Answer` and `ReasonContent` are populated
correctly
- `TestMistralChatHandlesStructuredContentWithoutThinking` —
`magistral-*` with a trivial answer returns a structured shape that has
only a `text` part; `ReasonContent` must stay empty
- `TestMistralChatIgnoresUnknownContentPartTypes` — `audio_url` and
`future_part_type` parts are skipped, `text` parts still flow through
- `TestExtractMistralContent` — table-driven unit coverage of the helper
for string, empty string, nil, empty array, text-only, thinking+text,
unsupported root type
```
$ go test -vet=off -run "TestMistral|TestExtractMistralContent" -count=1 -v ./internal/entity/models/...
=== RUN TestMistralChatHandlesStringContent
--- PASS: TestMistralChatHandlesStringContent (0.00s)
=== RUN TestMistralChatExtractsReasoningFromStructuredContent
--- PASS: TestMistralChatExtractsReasoningFromStructuredContent (0.00s)
=== RUN TestMistralChatHandlesStructuredContentWithoutThinking
--- PASS: TestMistralChatHandlesStructuredContentWithoutThinking (0.00s)
=== RUN TestMistralChatIgnoresUnknownContentPartTypes
--- PASS: TestMistralChatIgnoresUnknownContentPartTypes (0.00s)
=== RUN TestExtractMistralContent
=== RUN TestExtractMistralContent/plain_string
=== RUN TestExtractMistralContent/empty_string
=== RUN TestExtractMistralContent/nil
=== RUN TestExtractMistralContent/empty_array
=== RUN TestExtractMistralContent/text_only
=== RUN TestExtractMistralContent/thinking_then_text
=== RUN TestExtractMistralContent/unknown_root_type
--- PASS: TestExtractMistralContent (0.00s)
PASS
ok ragflow/internal/entity/models 0.046s
```
All 32 Mistral tests pass on go 1.25. `go build
./internal/entity/models/...` exits 0.
**Live integration test** — driver exercised against `api.mistral.ai/v1`
with the patched code:
```
=== RUN TestMistralMagistralSmoke
[OK] "magistral-small-latest" present upstream
[OK] "magistral-medium-latest" present upstream
[OK trivial] Answer="7" ReasonContent=""
[OK reasoning] Answer len=797 head="To determine when the two trains meet, we can follow these steps:\n\n1. **Identify..."
ReasonContent len=1069 head="Okay, let's see. There are two trains, one going 60 mph and the other going 90 mph. They're moving towards each other, s..."
MAGISTRAL SMOKE PASSED
--- PASS: TestMistralMagistralSmoke (18.09s)
PASS
ok ragflow/internal/entity/models 18.112s
```
What the live run proves on the wire:
- `magistral-small-latest` with a trivial prompt still uses the
string-content shape; the regression-net path is exercised against the
real server, not just the mock.
- `magistral-medium-latest` with a reasoning prompt uses the
structured-array shape; the new code path extracts a 1069-character
reasoning trace into `ChatResponse.ReasonContent` and a 797-character
visible answer into `ChatResponse.Answer`. Before this fix, the same
call returned `"invalid content format"` and the caller saw nothing.
The smoke-test file itself is not committed (live tests live outside the
PR diff, same convention used for prior provider PRs).
### Type of change
- [x] Bug Fix (non-breaking change which fixes an issue)
## What
- Add Perplexity as a chat and embedding provider backed by its
OpenAI-compatible `/chat/completions` and `/v1/embeddings` APIs
- Register Perplexity in the Go model factory and provider config
- Support non-streaming chat, SSE streaming chat, embeddings, model
listing, and connection checks
Refs #14736
---------
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
- Adds an `Astraflow` Go driver so the new API server can route
Astraflow (UCloud ModelVerse) chat instances, matching the existing
Python `AstraflowChat` (`rag/llm/chat_model.py:1237`). Follows the same
SaaS-driver shape used for Avian, Novita, TogetherAI, Replicate,
DeepInfra, Upstage, and LongCat.
Closes#15062
---------
Co-authored-by: Jin Hai <haijin.chn@gmail.com>
