Whatever message this page gives is out now! Go check it out!
Admin UI - AI Services
Last update:
May 18, 2026
ColdFusion 2025.0.08 adds a new AI Services section to the ColdFusion Administrator, providing a central place to configure LLM providers, embedding models, vector stores, RAG pipelines, and MCP servers without writing code.
Navigating to AI Services
- Log in to the ColdFusion Administrator.
- On the home page, click the AI Services tile.
- The AI Services section opens with the following tabs:
- LLMs: Configure LLM providers
- Embedding Models: Configure embedding model providers
- Vector Stores: Configure vector database connections
- RAG: Configure retrieval-augmented generation pipelines
- MCP Servers: Configure Model Context Protocol server connections
LLMs tab
The LLMs tab lets you add, view, edit, and delete LLM configurations. Each configuration is saved with a unique alias that can be referenced in your CFML code.
See LLMs for more information.
Add LLM Configuration
Fill in the following fields to register a new LLM provider configuration:
Identity fields
| Field | Required | Description |
|---|---|---|
| Config Alias | Yes | Unique identifier for this configuration. Alphanumeric, underscore, and hyphen characters only. Example: gpt4-prod, openaiconfig |
| Provider | Yes | Select the LLM provider from the dropdown. |
| Model Name | Conditional | Model identifier to use. Required for all providers except Azure OpenAI. |
Connection fields
| Field | Required | Description |
|---|---|---|
| API Key | Conditional | Authentication key for the provider API. Leave blank to supply at runtime. Not used by Ollama. |
| Endpoint URL | Conditional | Custom API endpoint URL. Required for Azure OpenAI and Ollama. |
| Deployment Name | Conditional | Azure OpenAI deployment name. Required for Azure only. |
| API Version | Conditional | Azure API version string (e.g., 2024-02-15-preview). Azure only. |
| Organization ID | Optional | OpenAI organization ID. OpenAI only. |
| Service Version | Optional | Azure OpenAI service version. Azure only. |
Model Parameters
Tune the model's behavior using the following optional parameters:
| Field | Description | Applicable providers |
|---|---|---|
| Temperature | Controls output randomness. Range: 0–2 typical. | All |
| Max Tokens | Maximum response length in tokens. | All |
| Top P | Nucleus sampling threshold (0–1 typical). | All |
| Top K | Number of top tokens to consider during sampling. | Gemini, Anthropic |
| Presence Penalty | Penalizes tokens based on appearance in text (>= 0). | OpenAI, Azure OpenAI |
| Repeat Penalty | Penalizes repeated tokens (>= 0). | Ollama |
| Seed | Random seed for deterministic outputs (> 0). | OpenAI, Azure OpenAI, Ollama |
| Response Format | Output format for model responses (Default or JSON). | OpenAI, Azure OpenAI, Ollama |
| Stop Sequences | Comma-separated strings that stop generation (e.g., \n\n, END). | All |
Advanced Settings
Additional settings are available under the Advanced Settings section of the Add LLM Configuration form, including HTTP client and proxy configuration.
Configured LLMs
Below the Add LLM Configuration form, the Configured LLMs table lists all saved configurations with the following columns:
| Column | Description |
|---|---|
| Actions | Edit or delete the configuration |
| Config Alias | The unique name given to the configuration |
| Provider | The selected LLM provider |
| Model | The model name |
| API Key | Masked display of the stored API key |
Note:
Disclaimer: The listed models/services are offerings of third-party providers supported for use with AI Services. Adobe does not own, control, or operate them. All names, trademarks, and associated branding are the property of their respective third-party owners and are used for identification purposes only. Their inclusion does not imply any partnership with, endorsement by, or sponsorship of Adobe.
<cfscript>
//add chat model config via admin ui/api/cfsetup with these 3 details , save config name as anth
//chatModelConfig = {
// PROVIDER: "anthropic",
// APIKEY: "#application.anthropicKey#",
// MODELNAME: "claude-sonnet-4-5"
// };
chatModel = ChatModel("anth");
response = chatModel.chat("What is 2 + 2?");
hasResponse = len(response.message ?: "") > 0;
containsFour = findNoCase("4", response.message ?: "") > 0 ||
findNoCase("four", response.message ?: "") > 0;
writeOutput("provider:anthropic | operation:basicChat | hasResponse:" & (hasResponse ? "YES" : "NO") & " | correctAnswer:" & (containsFour ? "YES" : "NO"));
agent = Agent({CHATMODEL : chatModel});
res = agent.chat("hello");
writeDump(res);
</cfscript>
Example 2
An OpenAI configuration named "myopenai" is saved, containing the provider and model name. Additional configuration parameters can be passed to override any existing values in the saved configuration, regardless of whether those parameters were originally defined.
<cfscript>
chatModel = ChatModel("myopenai", { apiKey : "#application.apiKey#" });
response = chatModel.chat("What is 2 + 3? how did to calculate it?");
writeDump(var=response);
</cfscript>
We're glad. Tell us how this page helped.
We're sorry. Can you tell us what didn't work for you?
Share additional feedback to help us improve.
0/255
|
