You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Commit f7042a4
Browse filesBrowse the repository at this point in the historyBrowse files
🤖 docs: Update the Endpoint and Agent librechat.yaml Reference (#863)
* docs: Update the endpoint and agent librechat.yaml reference
Check the agents, summarization, Azure, Vertex, Bedrock, assistants,
custom endpoint, model spec and shared endpoint pages, plus the AI
endpoint guides, against the LibreChat config schemas: fix wrong keys
and defaults, remove keys the server no longer reads and document the
missing ones.
* docs: Give the decision protocol rollout steps for conversation moves
Protocol version 1 is advertised when CODE_ENVIRONMENT_DECISION_VERSION
is unset, so a rolling upgrade must set it to 0 before the first new
replica starts and remove it once every replica is upgraded.
* docs: Correct the Azure Assistants version, machine choice and assistants flag
azureAssistants falls back to version 1 only when
endpoints.azureOpenAI.assistants is enabled. maxEnvironmentChoices
defaults to 32. Show the assistants flag on the group, where it takes
effect, rather than on a model.
Copy file name to clipboardExpand all lines: content/docs/configuration/librechat_yaml/ai_endpoints/azure.mdx
+1-40Lines changed: 1 addition & 40 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -17,7 +17,6 @@ endpoints:
17
17
azureOpenAI:
18
18
# Endpoint-level configuration
19
19
titleModel: "llama-70b-chat"
20
-
plugins: true
21
20
assistants: true
22
21
groups:
23
22
# Group-level configuration
@@ -94,15 +93,11 @@ To integrate Azure OpenAI with LibreChat, specific fields must be configured in
94
93
<OptionTable
95
94
options={[
96
95
['titleModel', 'string', 'Specifies the model to use for generating conversation titles. If not provided, the default model is set as `gpt-3.5-turbo`, which will result in no titles if lacking this model. You can also set this to dynamically use the current model by setting it to `current_model`.', 'titleModel:'],
97
-
['plugins', 'boolean', 'Enables the use of plugins through Azure. Set to `true` to activate Plugins endpoint support through your Azure config. Default: `false`.', 'plugins:false'],
98
96
['assistants', 'boolean', 'Enables the use of assistants through Azure. Set to `true` to activate Assistants endpoint through your Azure config. Default: `false`. Note: this requires an assistants-compatible region.', 'assistants:false'],
99
-
['summarize', 'boolean', 'Enables conversation summarization for all Azure models. Set to `true` to activate summarization. Default: `false`.', 'summarize:false'],
100
-
['summaryModel', 'string', 'Specifies the model to use for generating conversation summaries. If not provided, the default behavior is to use the first model in the `default` array of the first group.', 'summaryModel:'],
101
97
['titleConvo', 'boolean', 'Enables conversation title generation for all Azure models. Set to `true` to activate title generation. Default: `false`.', 'titleConvo:false'],
102
98
['titleMethod', 'string', 'Controls the method used for generating conversation titles. Valid values: "completion"(default), "structured", "functions" (legacy alias for "structured").', 'titleMethod:completion'],
103
99
['titlePrompt', 'string', 'Custom prompt for title generation. Must include {convo} placeholder for the conversation content.', 'See documentation for default prompt'],
104
100
['titlePromptTemplate', 'string', 'Template for formatting conversation content. Must include {input} and {output} placeholders. Default: "User: {input}\\nAI: {output}"', 'titlePromptTemplate:'],
105
-
['titleEndpoint', 'string', 'Alternative endpoint to use for title generation. Accepted values: openAI, azureOpenAI, google, anthropic, bedrock, or custom endpoint names.', 'titleEndpoint:'],
106
101
]}
107
102
/>
108
103
@@ -113,25 +108,13 @@ To integrate Azure OpenAI with LibreChat, specific fields must be configured in
113
108
]}
114
109
/>
115
110
116
-
**Custom Order (Optional):**
117
-
<OptionTable
118
-
options={[
119
-
['customOrder', 'number', 'Allows you to specify a custom order for the Azure endpoint in the user interface. Higher numbers will appear lower in the list. If not provided, the default order is determined by the order in which the endpoints are defined in the `librechat.yaml` file.', 'customOrder:'],
120
-
]}
121
-
/>
122
-
123
-
The `customOrder` option is commented out, as it is optional.
124
-
125
111
Example of these endpoint-level settings in your `librechat.yaml` file:
126
112
127
113
```yaml filename="librechat.yaml"
128
114
endpoints:
129
115
azureOpenAI:
130
116
titleModel: "gpt-3.5-turbo-1106"
131
-
plugins: true
132
117
assistants: true
133
-
summarize: true
134
-
summaryModel: "gpt-3.5-turbo-1106"
135
118
titleConvo: true
136
119
titleMethod: "functions"
137
120
groups:
@@ -358,8 +341,7 @@ endpoints:
358
341
- For this reason, it's recommended you use only one region or resource group for Azure OpenAI Assistants, or you will experience an error.
359
342
- Uploading to "OpenAI" is the default behavior for official `code_interpreter` and `retrieval` capabilities.
360
343
- Downloading files that assistants generate will soon be supported.
361
-
- As of May 19th 2024, retrieval and streaming are not yet supported through Azure OpenAI.
362
-
- To avoid any errors with retrieval while it's not supported, it's recommended to disable the capability altogether through the `azureAssistants` endpoint config:
344
+
- The default `azureAssistants` capabilities include `retrieval`. To turn it off, list the capabilities you want explicitly and omit it:
- By default, all capabilities, except retrieval, are enabled.
375
-
376
-
### Using Plugins with Azure
377
-
378
-
To use the Plugins endpoint with Azure OpenAI, you need a deployment supporting **[function calling](https://techcommunity.microsoft.com/t5/azure-ai-services-blog/function-calling-is-now-available-in-azure-openai-service/ba-p/3879241)**. Otherwise, set "Functions" off in the Agent settings. When you are not using "functions" mode, it's recommended to have "skip completion" off as well, which is a review step of what the agent generated.
379
-
380
-
To use Azure with the Plugins endpoint, make sure the field `plugins` is set to `true` in your Azure OpenAI endpoint config:
381
-
382
-
```yaml filename="librechat.yaml"
383
-
# Example Azure OpenAI Object Structure
384
-
endpoints:
385
-
azureOpenAI:
386
-
plugins: true # <------- Set this
387
-
groups:
388
-
# omitted for brevity
389
-
```
390
-
391
-
Configuring the `plugins` field will configure Plugins to use Azure models.
392
-
393
-
**NOTE**: The current configuration through `librechat.yaml` uses the primary model you select from the frontend for Plugin use, which is not usually how it works without Azure, where instead the "Agent" model is used. The Agent model setting can be ignored when using Plugins through Azure.
394
-
395
356
### Using a Specified Base URL with Azure
396
357
397
358
The base URL for Azure OpenAI API requests can be dynamically configured. This is useful for proxying services such as [Cloudflare AI Gateway](https://developers.cloudflare.com/ai-gateway/providers/azureopenai/), or if you wish to explicitly override the baseURL handling of the app.
0 commit comments