PluginInvokeError when transforming agent message: "read llm model failed" with Tongyi model #362

New Issue

Closed

opened 2026-02-16 10:19:08 -05:00 by yindo · 11 comments

yindo commented 2026-02-16 10:19:08 -05:00

Owner

Copy Link

Copy Source

Originally created by @zclcz on GitHub (Jun 12, 2025).

Self Checks

• 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 (我已阅读并同意 Language Policy).
• [FOR CHINESE USERS] 请务必使用英文提交 Issue，否则会被关闭。谢谢！:)
• Please do not modify this template :) and fill in all the required fields.

Dify version

1.4.1

Cloud or Self Hosted

Self Hosted (Docker)

Steps to reproduce

run failed: Failed to transform agent message: PluginInvokeError: {"args":{},"error_type":"Exception","message":"read llm model failed: request failed: [tongyi] Error: PluginInvokeError: {"args":{"description":"[models] Error: 'description'"},"error_type":"InvokeError","message":"[models] Error: 'description'"}"}

Additional issue: In the model vendor list, DeepSeek's models can be seen, but they are not visible in the agent plugin. Could this be due to the fact that the model plugins in these two places are not the same?

agent with mcp:

In the model vendor list, DeepSeek's models can be seen:

model plugin in agent: but they are not visible in the agent plugin

✔️ Expected Behavior

No response

❌ Actual Behavior

No response

Originally created by @zclcz on GitHub (Jun 12, 2025). ### Self Checks - [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 (我已阅读并同意 [Language Policy](https://github.com/langgenius/dify/issues/1542)). - [x] [FOR CHINESE USERS] 请务必使用英文提交 Issue，否则会被关闭。谢谢！:) - [x] Please do not modify this template :) and fill in all the required fields. ### Dify version 1.4.1 ### Cloud or Self Hosted Self Hosted (Docker) ### Steps to reproduce run failed: Failed to transform agent message: PluginInvokeError: {"args":{},"error_type":"Exception","message":"read llm model failed: request failed: [tongyi] Error: PluginInvokeError: {\"args\":{\"description\":\"[models] Error: 'description'\"},\"error_type\":\"InvokeError\",\"message\":\"[models] Error: 'description'\"}"} Additional issue: In the model vendor list, DeepSeek's models can be seen, but they are not visible in the agent plugin. Could this be due to the fact that the model plugins in these two places are not the same?  agent with mcp:  In the model vendor list, DeepSeek's models can be seen:   model plugin in agent: but they are not visible in the agent plugin  ### ✔️ Expected Behavior _No response_ ### ❌ Actual Behavior _No response_

yindo closed this issue 2026-02-16 10:19:08 -05:00

yindo commented 2026-02-16 10:19:08 -05:00

Author

Owner

Copy Link

Copy Source

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

Your PluginInvokeError with the Tongyi model in Docker is likely due to network or configuration issues. Make sure you are not using localhost or 127.0.0.1 for service endpoints inside Docker—use your host’s internal IP or http://host.docker.internal instead. Double-check that PLUGIN_DAEMON_URL is set correctly (e.g., http://host.docker.internal:5002), and that SANDBOX_ENABLE_NETWORK=true is set in your environment. Also, confirm your Tongyi DashScope API key is present in your environment variables and that the Tongyi plugin is properly installed and models are listed in the Dify admin interface. If the plugin is outdated, upgrade it, as older versions may not support new models or have registration bugs. Sometimes, models need to be added manually or you may need to wait for a plugin update to get new models built-in [source].

For the DeepSeek visibility issue: DeepSeek is a built-in (static) provider, not a plugin-based provider. The agent plugin infrastructure only exposes plugin-based models/strategies, so DeepSeek appears in the vendor list but not in the agent plugin selection. Only models that support tool-calling (function call or ReACT) are visible in the Agent node. If you have unconfigured model plugins, removing or disabling them may help with incomplete listings [source].

If you continue to see validation errors, ensure your model configuration matches the expected schema and that the model parameter is correctly defined in your workflow. Switching to a different model (like DeepSeek) has resolved similar errors for other users [source].

To reply, just mention @dosu.

---

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

@dosubot[bot] commented on GitHub (Jun 12, 2025): <!-- Answer --> Your PluginInvokeError with the Tongyi model in Docker is likely due to network or configuration issues. Make sure you are not using `localhost` or `127.0.0.1` for service endpoints inside Docker—use your host’s internal IP or `http://host.docker.internal` instead. Double-check that `PLUGIN_DAEMON_URL` is set correctly (e.g., `http://host.docker.internal:5002`), and that `SANDBOX_ENABLE_NETWORK=true` is set in your environment. Also, confirm your Tongyi DashScope API key is present in your environment variables and that the Tongyi plugin is properly installed and models are listed in the Dify admin interface. If the plugin is outdated, upgrade it, as older versions may not support new models or have registration bugs. Sometimes, models need to be added manually or you may need to wait for a plugin update to get new models built-in [[source](https://github.com/langgenius/dify/issues/20732)]. For the DeepSeek visibility issue: DeepSeek is a built-in (static) provider, not a plugin-based provider. The agent plugin infrastructure only exposes plugin-based models/strategies, so DeepSeek appears in the vendor list but not in the agent plugin selection. Only models that support tool-calling (function call or ReACT) are visible in the Agent node. If you have unconfigured model plugins, removing or disabling them may help with incomplete listings [[source](https://github.com/langgenius/dify/issues/18335)]. If you continue to see validation errors, ensure your model configuration matches the expected schema and that the model parameter is correctly defined in your workflow. Switching to a different model (like DeepSeek) has resolved similar errors for other users [[source](https://github.com/langgenius/dify/issues/14684)]. <!-- Dosu Comment Footer --> *To reply, just mention [@dosu](https://go.dosu.dev/dosubot).* --- ^{How did I do? [Good](https://app.dosu.dev/response-feedback/0134ff86-4d0d-4166-9bc6-566c4b57d9d3?feedback_type=great_response) | [Irrelevant](https://app.dosu.dev/response-feedback/0134ff86-4d0d-4166-9bc6-566c4b57d9d3?feedback_type=irrelevant_answer) | [Incorrect](https://app.dosu.dev/response-feedback/0134ff86-4d0d-4166-9bc6-566c4b57d9d3?feedback_type=incorrect_sources) | [Verbose](https://app.dosu.dev/response-feedback/0134ff86-4d0d-4166-9bc6-566c4b57d9d3?feedback_type=too_verbose) | [Hallucination](https://app.dosu.dev/response-feedback/0134ff86-4d0d-4166-9bc6-566c4b57d9d3?feedback_type=hallucination) | [Report 🐛](https://app.dosu.dev/response-feedback/0134ff86-4d0d-4166-9bc6-566c4b57d9d3?feedback_type=bug_report) | [Other](https://app.dosu.dev/response-feedback/0134ff86-4d0d-4166-9bc6-566c4b57d9d3?feedback_type=other)}&nbsp;&nbsp;[](https://go.dosu.dev/discord-bot)&nbsp;[](https://twitter.com/intent/tweet?text=%40dosu_ai%20helped%20me%20solve%20this%20issue!&url=https%3A//github.com/langgenius/dify/issues/20946)

yindo commented 2026-02-16 10:19:08 -05:00

Author

Owner

Copy Link

Copy Source

@zclcz commented on GitHub (Jun 12, 2025):

Your PluginInvokeError with the Tongyi model in Docker is likely due to network or configuration issues. Make sure you are not using localhost or 127.0.0.1 for service endpoints inside Docker—use your host’s internal IP or http://host.docker.internal instead. Double-check that PLUGIN_DAEMON_URL is set correctly (e.g., http://host.docker.internal:5002), and that SANDBOX_ENABLE_NETWORK=true is set in your environment. Also, confirm your Tongyi DashScope API key is present in your environment variables and that the Tongyi plugin is properly installed and models are listed in the Dify admin interface. If the plugin is outdated, upgrade it, as older versions may not support new models or have registration bugs. Sometimes, models need to be added manually or you may need to wait for a plugin update to get new models built-in [source].

For the DeepSeek visibility issue: DeepSeek is a built-in (static) provider, not a plugin-based provider. The agent plugin infrastructure only exposes plugin-based models/strategies, so DeepSeek appears in the vendor list but not in the agent plugin selection. Only models that support tool-calling (function call or ReACT) are visible in the Agent node. If you have unconfigured model plugins, removing or disabling them may help with incomplete listings [source].

If you continue to see validation errors, ensure your model configuration matches the expected schema and that the model parameter is correctly defined in your workflow. Switching to a different model (like DeepSeek) has resolved similar errors for other users [source].

To reply, just mention @dosu.

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

I can use Qwen3 normally in other workflows, but it doesn't work here. I've ruled out network and configuration issues, so the problem likely lies within the agent node. Could you please check if this version has this issue? Thank you.

@zclcz commented on GitHub (Jun 12, 2025): > Your PluginInvokeError with the Tongyi model in Docker is likely due to network or configuration issues. Make sure you are not using `localhost` or `127.0.0.1` for service endpoints inside Docker—use your host’s internal IP or `http://host.docker.internal` instead. Double-check that `PLUGIN_DAEMON_URL` is set correctly (e.g., `http://host.docker.internal:5002`), and that `SANDBOX_ENABLE_NETWORK=true` is set in your environment. Also, confirm your Tongyi DashScope API key is present in your environment variables and that the Tongyi plugin is properly installed and models are listed in the Dify admin interface. If the plugin is outdated, upgrade it, as older versions may not support new models or have registration bugs. Sometimes, models need to be added manually or you may need to wait for a plugin update to get new models built-in [[source](https://github.com/langgenius/dify/issues/20732)]. > > For the DeepSeek visibility issue: DeepSeek is a built-in (static) provider, not a plugin-based provider. The agent plugin infrastructure only exposes plugin-based models/strategies, so DeepSeek appears in the vendor list but not in the agent plugin selection. Only models that support tool-calling (function call or ReACT) are visible in the Agent node. If you have unconfigured model plugins, removing or disabling them may help with incomplete listings [[source](https://github.com/langgenius/dify/issues/18335)]. > > If you continue to see validation errors, ensure your model configuration matches the expected schema and that the model parameter is correctly defined in your workflow. Switching to a different model (like DeepSeek) has resolved similar errors for other users [[source](https://github.com/langgenius/dify/issues/14684)]. > > _To reply, just mention [@dosu](https://go.dosu.dev/dosubot)._ > > How did I do? [Good](https://app.dosu.dev/response-feedback/0134ff86-4d0d-4166-9bc6-566c4b57d9d3?feedback_type=great_response) | [Irrelevant](https://app.dosu.dev/response-feedback/0134ff86-4d0d-4166-9bc6-566c4b57d9d3?feedback_type=irrelevant_answer) | [Incorrect](https://app.dosu.dev/response-feedback/0134ff86-4d0d-4166-9bc6-566c4b57d9d3?feedback_type=incorrect_sources) | [Verbose](https://app.dosu.dev/response-feedback/0134ff86-4d0d-4166-9bc6-566c4b57d9d3?feedback_type=too_verbose) | [Hallucination](https://app.dosu.dev/response-feedback/0134ff86-4d0d-4166-9bc6-566c4b57d9d3?feedback_type=hallucination) | [Report 🐛](https://app.dosu.dev/response-feedback/0134ff86-4d0d-4166-9bc6-566c4b57d9d3?feedback_type=bug_report) | [Other](https://app.dosu.dev/response-feedback/0134ff86-4d0d-4166-9bc6-566c4b57d9d3?feedback_type=other)  [](https://go.dosu.dev/discord-bot) [](https://twitter.com/intent/tweet?text=%40dosu_ai%20helped%20me%20solve%20this%20issue!&url=https%3A//github.com/langgenius/dify/issues/20946) I can use Qwen3 normally in other workflows, but it doesn't work here. I've ruled out network and configuration issues, so the problem likely lies within the agent node. Could you please check if this version has this issue? Thank you.

yindo commented 2026-02-16 10:19:08 -05:00

Author

Owner

Copy Link

Copy Source

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

Based on everything I found, there’s no evidence of a general bug in Dify 1.4.1 that prevents Qwen3 from working in the agent node, but there are a few known causes for similar PluginInvokeError issues:

• The Qwen3 model requires the Tongyi plugin version 0.0.19 or higher. The plugin must be up to date, and the stream parameter is handled automatically—no manual config needed for that part. If your plugin is older, upgrade it and restart the plugin daemon container to ensure compatibility and proper model registration.
• Models only appear in the Agent node if they support tool-calling (function call or ReACT). If Qwen3 is missing, check that your plugin version exposes these capabilities and that there are no unconfigured or outdated plugins in your provider list. Removing or updating those can resolve visibility issues [source].
• If you see validation errors or PluginInvokeError mentioning FunctionCallingParams or model.entity, double-check that your workflow’s model configuration matches the schema expected by Dify (all required fields, correct types). This is a common cause of agent node failures, even if the model works elsewhere [source].
• There are no recent changes in 1.4.1 that would break Qwen3 in the agent node specifically. The agent node is model-agnostic and relies on the plugin/provider system for model integration, so most issues are at the plugin or configuration level, not the core code.

If you’ve confirmed your plugin version and configuration are correct, and Qwen3 still works in other nodes but not the agent node, try updating the Tongyi plugin, removing any unconfigured plugins, and ensuring your workflow’s model config is complete. If the issue persists, it may be a plugin compatibility edge case, but there’s no general blocker in 1.4.1 for Qwen3 agent node use [source].

Let me know if you want more details on any of these checks!

To reply, just mention @dosu.

---

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

@dosubot[bot] commented on GitHub (Jun 12, 2025): <!-- Answer --> Based on everything I found, there’s no evidence of a general bug in Dify 1.4.1 that prevents Qwen3 from working in the agent node, but there are a few known causes for similar PluginInvokeError issues: - The Qwen3 model requires the Tongyi plugin version 0.0.19 or higher. The plugin must be up to date, and the stream parameter is handled automatically—no manual config needed for that part. If your plugin is older, upgrade it and restart the plugin daemon container to ensure compatibility and proper model registration. - Models only appear in the Agent node if they support tool-calling (function call or ReACT). If Qwen3 is missing, check that your plugin version exposes these capabilities and that there are no unconfigured or outdated plugins in your provider list. Removing or updating those can resolve visibility issues [[source](https://github.com/langgenius/dify/issues/18335)]. - If you see validation errors or PluginInvokeError mentioning FunctionCallingParams or model.entity, double-check that your workflow’s model configuration matches the schema expected by Dify (all required fields, correct types). This is a common cause of agent node failures, even if the model works elsewhere [[source](https://github.com/langgenius/dify/issues/14684)]. - There are no recent changes in 1.4.1 that would break Qwen3 in the agent node specifically. The agent node is model-agnostic and relies on the plugin/provider system for model integration, so most issues are at the plugin or configuration level, not the core code. If you’ve confirmed your plugin version and configuration are correct, and Qwen3 still works in other nodes but not the agent node, try updating the Tongyi plugin, removing any unconfigured plugins, and ensuring your workflow’s model config is complete. If the issue persists, it may be a plugin compatibility edge case, but there’s no general blocker in 1.4.1 for Qwen3 agent node use [[source](https://github.com/langgenius/dify/issues/19126)]. Let me know if you want more details on any of these checks! <!-- Dosu Comment Footer --> *To reply, just mention [@dosu](https://go.dosu.dev/dosubot).* --- ^{How did I do? [Good](https://app.dosu.dev/response-feedback/7d1f75d6-5ead-4c8c-befa-218a01cf2663?feedback_type=great_response) | [Irrelevant](https://app.dosu.dev/response-feedback/7d1f75d6-5ead-4c8c-befa-218a01cf2663?feedback_type=irrelevant_answer) | [Incorrect](https://app.dosu.dev/response-feedback/7d1f75d6-5ead-4c8c-befa-218a01cf2663?feedback_type=incorrect_sources) | [Verbose](https://app.dosu.dev/response-feedback/7d1f75d6-5ead-4c8c-befa-218a01cf2663?feedback_type=too_verbose) | [Hallucination](https://app.dosu.dev/response-feedback/7d1f75d6-5ead-4c8c-befa-218a01cf2663?feedback_type=hallucination) | [Report 🐛](https://app.dosu.dev/response-feedback/7d1f75d6-5ead-4c8c-befa-218a01cf2663?feedback_type=bug_report) | [Other](https://app.dosu.dev/response-feedback/7d1f75d6-5ead-4c8c-befa-218a01cf2663?feedback_type=other)}&nbsp;&nbsp;[](https://go.dosu.dev/discord-bot)&nbsp;[](https://twitter.com/intent/tweet?text=%40dosu_ai%20helped%20me%20solve%20this%20issue!&url=https%3A//github.com/langgenius/dify/issues/20946)

yindo commented 2026-02-16 10:19:08 -05:00

Author

Owner

Copy Link

Copy Source

@zclcz commented on GitHub (Jun 12, 2025):

Based on everything I found, there’s no evidence of a general bug in Dify 1.4.1 that prevents Qwen3 from working in the agent node, but there are a few known causes for similar PluginInvokeError issues:

• The Qwen3 model requires the Tongyi plugin version 0.0.19 or higher. The plugin must be up to date, and the stream parameter is handled automatically—no manual config needed for that part. If your plugin is older, upgrade it and restart the plugin daemon container to ensure compatibility and proper model registration.
• Models only appear in the Agent node if they support tool-calling (function call or ReACT). If Qwen3 is missing, check that your plugin version exposes these capabilities and that there are no unconfigured or outdated plugins in your provider list. Removing or updating those can resolve visibility issues [source].
• If you see validation errors or PluginInvokeError mentioning FunctionCallingParams or model.entity, double-check that your workflow’s model configuration matches the schema expected by Dify (all required fields, correct types). This is a common cause of agent node failures, even if the model works elsewhere [source].
• There are no recent changes in 1.4.1 that would break Qwen3 in the agent node specifically. The agent node is model-agnostic and relies on the plugin/provider system for model integration, so most issues are at the plugin or configuration level, not the core code.

If you’ve confirmed your plugin version and configuration are correct, and Qwen3 still works in other nodes but not the agent node, try updating the Tongyi plugin, removing any unconfigured plugins, and ensuring your workflow’s model config is complete. If the issue persists, it may be a plugin compatibility edge case, but there’s no general blocker in 1.4.1 for Qwen3 agent node use [source].

Let me know if you want more details on any of these checks!

To reply, just mention @dosu.

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

My Tongyi plugin is only 0.1 version behind — it's almost up to date. Moreover, this isn't just an issue with selecting the Qwen3 model; the same error occurs when selecting any model. Therefore, the problem is definitely with the "agent" node.

@zclcz commented on GitHub (Jun 12, 2025): > Based on everything I found, there’s no evidence of a general bug in Dify 1.4.1 that prevents Qwen3 from working in the agent node, but there are a few known causes for similar PluginInvokeError issues: > > * The Qwen3 model requires the Tongyi plugin version 0.0.19 or higher. The plugin must be up to date, and the stream parameter is handled automatically—no manual config needed for that part. If your plugin is older, upgrade it and restart the plugin daemon container to ensure compatibility and proper model registration. > * Models only appear in the Agent node if they support tool-calling (function call or ReACT). If Qwen3 is missing, check that your plugin version exposes these capabilities and that there are no unconfigured or outdated plugins in your provider list. Removing or updating those can resolve visibility issues [[source](https://github.com/langgenius/dify/issues/18335)]. > * If you see validation errors or PluginInvokeError mentioning FunctionCallingParams or model.entity, double-check that your workflow’s model configuration matches the schema expected by Dify (all required fields, correct types). This is a common cause of agent node failures, even if the model works elsewhere [[source](https://github.com/langgenius/dify/issues/14684)]. > * There are no recent changes in 1.4.1 that would break Qwen3 in the agent node specifically. The agent node is model-agnostic and relies on the plugin/provider system for model integration, so most issues are at the plugin or configuration level, not the core code. > > If you’ve confirmed your plugin version and configuration are correct, and Qwen3 still works in other nodes but not the agent node, try updating the Tongyi plugin, removing any unconfigured plugins, and ensuring your workflow’s model config is complete. If the issue persists, it may be a plugin compatibility edge case, but there’s no general blocker in 1.4.1 for Qwen3 agent node use [[source](https://github.com/langgenius/dify/issues/19126)]. > > Let me know if you want more details on any of these checks! > > _To reply, just mention [@dosu](https://go.dosu.dev/dosubot)._ > > How did I do? [Good](https://app.dosu.dev/response-feedback/7d1f75d6-5ead-4c8c-befa-218a01cf2663?feedback_type=great_response) | [Irrelevant](https://app.dosu.dev/response-feedback/7d1f75d6-5ead-4c8c-befa-218a01cf2663?feedback_type=irrelevant_answer) | [Incorrect](https://app.dosu.dev/response-feedback/7d1f75d6-5ead-4c8c-befa-218a01cf2663?feedback_type=incorrect_sources) | [Verbose](https://app.dosu.dev/response-feedback/7d1f75d6-5ead-4c8c-befa-218a01cf2663?feedback_type=too_verbose) | [Hallucination](https://app.dosu.dev/response-feedback/7d1f75d6-5ead-4c8c-befa-218a01cf2663?feedback_type=hallucination) | [Report 🐛](https://app.dosu.dev/response-feedback/7d1f75d6-5ead-4c8c-befa-218a01cf2663?feedback_type=bug_report) | [Other](https://app.dosu.dev/response-feedback/7d1f75d6-5ead-4c8c-befa-218a01cf2663?feedback_type=other)  [](https://go.dosu.dev/discord-bot) [](https://twitter.com/intent/tweet?text=%40dosu_ai%20helped%20me%20solve%20this%20issue!&url=https%3A//github.com/langgenius/dify/issues/20946) My Tongyi plugin is only 0.1 version behind — it's almost up to date. Moreover, this isn't just an issue with selecting the Qwen3 model; the same error occurs when selecting any model. Therefore, the problem is definitely with the "agent" node.

yindo commented 2026-02-16 10:19:08 -05:00

Author

Owner

Copy Link

Copy Source

@liaofuyan commented on GitHub (Jun 14, 2025):

碰到同样的问题

@liaofuyan commented on GitHub (Jun 14, 2025): 碰到同样的问题

yindo commented 2026-02-16 10:19:08 -05:00

Author

Owner

Copy Link

Copy Source

@wanghan5 commented on GitHub (Jun 16, 2025):

I meet the same problem here when I'm trying to interact with MCP server.
Important clue: mcp works well in ReAct mode, but in functioncalling mode, the log in plugin-daemon pod shows that dify got the tools list and then failed in the next step.

@wanghan5 commented on GitHub (Jun 16, 2025): I meet the same problem here when I'm trying to interact with MCP server. Important clue: mcp works well in ReAct mode, but in functioncalling mode, the log in plugin-daemon pod shows that dify got the tools list and then failed in the next step.

yindo commented 2026-02-16 10:19:08 -05:00

Author

Owner

Copy Link

Copy Source

@junjiem commented on GitHub (Jun 16, 2025):

see https://github.com/junjiem/dify-plugin-agent-mcp_sse/pull/89

@junjiem commented on GitHub (Jun 16, 2025): see https://github.com/junjiem/dify-plugin-agent-mcp_sse/pull/89

yindo commented 2026-02-16 10:19:08 -05:00

Author

Owner

Copy Link

Copy Source

@wanghan5 commented on GitHub (Jun 16, 2025):

see junjiem/dify-plugin-agent-mcp_sse#89

I think I have found the cause: the official tongyi plugin is not compatible with MCP, it will try to get description field in each properties . @junjiem @liaofuyan @zclcz

It works well when using openai-api plugin

@wanghan5 commented on GitHub (Jun 16, 2025): > see [junjiem/dify-plugin-agent-mcp_sse#89](https://github.com/junjiem/dify-plugin-agent-mcp_sse/pull/89) I think I have found the cause: the official tongyi plugin is not compatible with MCP, it will try to get description field in each properties . @junjiem @liaofuyan @zclcz It works well when using openai-api plugin 

yindo commented 2026-02-16 10:19:08 -05:00

Author

Owner

Copy Link

Copy Source

@zclcz commented on GitHub (Jun 17, 2025):

see junjiem/dify-plugin-agent-mcp_sse#89

I think I have found the cause: the official tongyi plugin is not compatible with MCP, it will try to get description field in each properties . @junjiem @liaofuyan @zclcz

It works well when using openai-api plugin

However, the Qwen model from the Tongyi provider works normally when using MCP in the 'Agent' feature.

@zclcz commented on GitHub (Jun 17, 2025): > > see [junjiem/dify-plugin-agent-mcp_sse#89](https://github.com/junjiem/dify-plugin-agent-mcp_sse/pull/89) > > I think I have found the cause: the official tongyi plugin is not compatible with MCP, it will try to get description field in each properties . [@junjiem](https://github.com/junjiem) [@liaofuyan](https://github.com/liaofuyan) [@zclcz](https://github.com/zclcz) > > It works well when using openai-api plugin > >  However, the Qwen model from the Tongyi provider works normally when using MCP in the 'Agent' feature. 

yindo commented 2026-02-16 10:19:08 -05:00

Author

Owner

Copy Link

Copy Source

@wanghan5 commented on GitHub (Jun 17, 2025):

see junjiem/dify-plugin-agent-mcp_sse#89

I think I have found the cause: the official tongyi plugin is not compatible with MCP, it will try to get description field in each properties . @junjiem @liaofuyan @zclcz
It works well when using openai-api plugin

However, the Qwen model from the Tongyi provider works normally when using MCP in the 'Agent' feature.

In "Agent", MCP Server is wrapped as a tool. It is not directly connectted to Agent

@wanghan5 commented on GitHub (Jun 17, 2025): > > > see [junjiem/dify-plugin-agent-mcp_sse#89](https://github.com/junjiem/dify-plugin-agent-mcp_sse/pull/89) > > > > > > I think I have found the cause: the official tongyi plugin is not compatible with MCP, it will try to get description field in each properties . [@junjiem](https://github.com/junjiem) [@liaofuyan](https://github.com/liaofuyan) [@zclcz](https://github.com/zclcz) > > It works well when using openai-api plugin > >  > > However, the Qwen model from the Tongyi provider works normally when using MCP in the 'Agent' feature.  In "Agent", MCP Server is wrapped as a tool. It is not directly connectted to Agent

yindo commented 2026-02-16 10:19:08 -05:00

Author

Owner

Copy Link

Copy Source

@dosubot[bot] commented on GitHub (Aug 21, 2025):

Hi, @zclcz. I'm Dosu, and I'm helping the dify-official-plugins team manage their backlog and am marking this issue as stale.

Issue Summary:

• You reported a PluginInvokeError "read llm model failed" when using the Tongyi model on Dify 1.4.1 self-hosted via Docker.
• DeepSeek models appear in the vendor list but not in the agent plugin interface.
• I suggested checking network settings, plugin versions (Tongyi plugin 0.0.19+ required), and model configurations.
• Compatibility issues exist between the official Tongyi plugin and MCP server; MCP works better via the openai-api plugin.
• The Qwen model from Tongyi works normally with MCP in the agent feature, indicating the issue is likely plugin compatibility or configuration-related.

Next Steps:

• Please confirm if this issue is still relevant with the latest version of the dify-official-plugins repository and updated plugin configurations.
• If it is, you can keep the discussion open by commenting here; otherwise, this issue will be automatically closed in 5 days.

Thank you for your understanding and contribution!

@dosubot[bot] commented on GitHub (Aug 21, 2025): Hi, @zclcz. I'm [Dosu](https://dosu.dev), and I'm helping the dify-official-plugins team manage their backlog and am marking this issue as stale. **Issue Summary:** - You reported a PluginInvokeError "read llm model failed" when using the Tongyi model on Dify 1.4.1 self-hosted via Docker. - DeepSeek models appear in the vendor list but not in the agent plugin interface. - I suggested checking network settings, plugin versions (Tongyi plugin 0.0.19+ required), and model configurations. - Compatibility issues exist between the official Tongyi plugin and MCP server; MCP works better via the openai-api plugin. - The Qwen model from Tongyi works normally with MCP in the agent feature, indicating the issue is likely plugin compatibility or configuration-related. **Next Steps:** - Please confirm if this issue is still relevant with the latest version of the dify-official-plugins repository and updated plugin configurations. - If it is, you can keep the discussion open by commenting here; otherwise, this issue will be automatically closed in 5 days. Thank you for your understanding and contribution!

yindo referenced this issue 2026-02-16 10:22:37 -05:00

[PR #362] [MERGED] Fix failed to create rerank model when use vllm rerank #1317

Sign in to join this conversation.

No Branch/Tag Specified

Branches Tags

main

feat/3479-gemini-flash-models

dependabot/uv/tools/chart/uv-cdf8d38825

feat/moonshot-kimi-k3-k27-code

feat/gemini-system-instruction-semantics

fix/gemini-text-system-parts

dosu/doc-update-3436

feat/openai-gpt-5-6-support

fix/outlook-trigger-readme-permissions

fix/microsoft-oauth-datasource-tenant

fix/byteplus-api-base-url-label

feat/ark-video-generation-custom-models

fix/openai-stream-reasoning-thinking

fix/skip-live-tests-missing-keys

codex/enhance-discord-integrations

codex/harden-trigger-oauth-permissions

update-lemonade-default-port

codex/fix-pr-review-ci

feat/ark-multimodal-polling

copilot/update-mineru-3x-plugin-file-types

feat/action-security-fix

feat/claude-fable5

codex/fix-ollama-model-category

codex/upgrade-ollama-latest

feat/minimax-m3

codex/refine-plugin-readmes

fix/fireworks-model-label-zh-hans

chore/version-bump-plugins-batch-06

chore/version-bump-plugins-batch-05

chore/version-bump-plugins-batch-04

chore/version-bump-plugins-batch-03

chore/version-bump-plugins-batch-02

fix/plugin-sdk-i18n-fields

chore/version-bump-plugins-batch-01

chore/dependencies-high-priority

chore/dependencies-medium-priority

chore/dependencies-low-priority-clean

chore/dependencies-low-priority

fix/tongyi-align-deepseek-model-metadata

codex/update-openrouter-models

codex/azure-ai-studio-cjk-version-bump

feat/ernie-image-tool

claude/cranky-thompson-b6a281

fix/remove-unavailable-volcengine-models

fix/legacy-plugin-daemon-compatibility

copilot/add-deepseek-v4-to-openrouter

codex/merge-tongyi-tts-prs

codex/depot-runners

chore/migrate-selected-plugins-to-uv-lock

fix/repair-failed-plugin-deploys

ci/fix-plugin-change-detection

chore/migrate-aws-s3-storage-to-uv-lock

chore/migrate-somark-dependencies-to-uv-lock

ci/uv-lock-dependency-validation

fix/2911-openai-api-compatible-tool-call-content

chore/remove-deprecated-plugins

copilot/add-more-models-openrouter-provider

copilot/fix-error-connecting-fireworks-ai

fix/openai-compatible-max-completion-tokens-validation

fix/gemini-lockfile

chore/add-React-fallback-desc

feat/add-gemini-embedding-2-preview

copilot/add-claude-sonnet-opus-4-6

codex/update-xai-models-2721

fix/remove-sentence-transformer-for-hubspot

feat/add-gpt-5.4

copilot/fix-ci-errors-and-bump-version

copilot/fix-security-errors-combined

copilot/update-gemini-model-version

fix/anthropic-parallel-tool-call

codex/fix-azure-ai-studio-azure-core-1-38

codex/fix-azure-blob-azure-core-1-38

codex/fix-azure-openai-azure-core-1-38

gitlab-credentials-c

fix/openrouter-validation-model

fix/gemini-null-checks

fix/deepseek-tool-message-ignore

fix/gemini-dependencies

No results found.

Labels

Clear labels

bug

bug

bug

dependencies

documentation

enhancement

good first issue

invalid

pull-request

Mirrored from GitHub Pull Request

question

stale

No Label

Milestone

No items

No Milestone

Projects

Clear projects

No project

Assignees

Clear assignees

yindo

No Assignees

1 Participants

Notifications

Subscribe

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: langgenius/dify-official-plugins#362