Overview
IBM watsonx is IBM's enterprise AI platform – text generation, embeddings, custom model deployments, prompt templates. The Automize watsonx connector covers AI inference + deployment operations.
- Regions: IBM Cloud regions.
- Plans: Per-model + per-RU pricing.
Setting up the connection
Auth type: Bearer Token.
- IBM Cloud → IAM → API key.
- Exchange API key for IAM bearer token (connector handles).
- In Automize → Settings → Connectors → IBM watsonx. Paste API key + project ID.
Find your credentials at https://cloud.ibm.com/iam/apikeys.
Rate limits
Per-instance + per-model.
Data model
AI primitives:
Project ─┬─ Models (foundation + custom)
├─ Deployments
├─ Prompt Templates
└─ Embeddings indexAvailable operations
16 operations available. Click any row to jump to its detail.
| Operation | Category | What it does |
|---|---|---|
| Cancel Training Job | General | Cancel a training job |
| Create Deployment | General | Create a model deployment |
| Create Prompt Template | General | Create a prompt template |
| Create Training Job | General | Create a training/tuning job |
| Delete Deployment | General | Delete a deployment |
| Generate Embeddings | General | Generate text embeddings |
| Generate Text | General | Generate text from a foundation model |
| Generate Text (Stream) | General | Generate text with streaming response |
| Get Deployment | General | Get deployment details |
| Get Model | General | Get model details |
| Get Training Job | General | Get training job details |
| List Deployments | General | List model deployments |
| List Models | General | List available foundation models |
| List Prompt Templates | General | List saved prompt templates |
| List Training Jobs | General | List training/tuning jobs |
| Tokenize | General | Tokenize text for a model |
Operations
Cancel Training Job
Cancel a training job
| Parameter | Required | Type | Description |
|---|---|---|---|
Training Id training_id | Yes | text |
Create Deployment
Create a model deployment
No input parameters.
Create Prompt Template
Create a prompt template
No input parameters.
Create Training Job
Create a training/tuning job
No input parameters.
Delete Deployment
Delete a deployment
| Parameter | Required | Type | Description |
|---|---|---|---|
Deployment Id deployment_id | Yes | text |
Generate Embeddings
Generate text embeddings
| Parameter | Required | Type | Description |
|---|---|---|---|
Model Id model_id | No | text | |
Inputs inputs | No | text | |
Project Id project_id | No | text |
Generate Text
Generate text from a foundation model
| Parameter | Required | Type | Description |
|---|---|---|---|
Model Id model_id | No | text | |
Input input | No | text | |
Parameters parameters | No | text | |
Project Id project_id | No | text |
Generate Text (Stream)
Generate text with streaming response
| Parameter | Required | Type | Description |
|---|---|---|---|
Model Id model_id | No | text | |
Input input | No | text | |
Parameters parameters | No | text | |
Project Id project_id | No | text |
Get Deployment
Get deployment details
| Parameter | Required | Type | Description |
|---|---|---|---|
Deployment Id deployment_id | Yes | text |
Get Model
Get model details
| Parameter | Required | Type | Description |
|---|---|---|---|
Model Id model_id | Yes | text |
Get Training Job
Get training job details
| Parameter | Required | Type | Description |
|---|---|---|---|
Training Id training_id | Yes | text |
List Deployments
List model deployments
No input parameters.
List Models
List available foundation models
No input parameters.
Tips
- Returns foundation models + your custom-trained deployments.
List Prompt Templates
List saved prompt templates
No input parameters.
Tips
- Reusable prompt configurations.
List Training Jobs
List training/tuning jobs
No input parameters.
Tokenize
Tokenize text for a model
| Parameter | Required | Type | Description |
|---|---|---|---|
Model Id model_id | No | text | |
Input input | No | text | |
Parameters parameters | No | text | |
Project Id project_id | No | text |
FAQ
- vs OpenAI / Anthropic?
- watsonx: IBM-supported, enterprise-grade, on-prem option, multiple model providers (Meta Llama, Mistral, IBM Granite). OpenAI/Anthropic: best-in-class quality on flagship models.