Skip to main content

​
Overview

Targets define the configuration for stimulus generation (stim) and trajectory generation (rollout). They are composed from a base template plus per-target overrides. Related Endpoints:

​
Endpoints

​
List Targets

List all targets in a repository. Endpoint: GET /v1/repo/{repo_name}/targets Request: 
curl -X GET "https://bapi.bread.com.ai/v1/repo/my_repo/targets" \
  -H "Authorization: Bearer $BREAD_API_KEY"
​
repo_name
string
required
Repository name
Response: 200 OK 
{
  "targets": ["target_1", "target_2", "coding_target"]
}

​
Get Target

Get target definition and metadata. Endpoint: GET /v1/repo/{repo_name}/targets/{target_name} Request: 
curl -X GET "https://bapi.bread.com.ai/v1/repo/my_repo/targets/coding_target" \
  -H "Authorization: Bearer $BREAD_API_KEY"
​
target_name
string
required
Target name
​
repo_name
string
required
Repository name
Response: 200 OK 
{
  "target_name": "coding_target",
  "config": {
    "model_name": "Qwen/Qwen3-32B",
    "teacher_prompt": "system_prompt_coding",
    "student_prompt": "user_prompt_python",
    "generators": [],
    "max_tokens": 150,
    "temperature": 1.0
  }
}

​
Create or Update Target

Create or update a target. Targets are composed from a base template plus per-target overrides. Endpoint: POST /v1/repo/{repo_name}/targets Request: 
curl -X POST "https://bapi.bread.com.ai/v1/repo/my_repo/targets" \
  -H "Authorization: Bearer $BREAD_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "target_name": "coding_target",
    "template": "default",
    "overrides": {
        "model_name": "Qwen/Qwen3-32B",
      "teacher_prompt": "system_prompt_coding",
      "student_prompt": "user_prompt_python",
        "generators": [
            {
                "type": "oneshot_qs",
                "model": "claude-sonnet-4-5-20250929",
                "numq": 10,
                "temperature": 1.0
            }
        ],
        "max_tokens": 150,
      "temperature": 1.0
    }
  }'
​
repo_name
string
required
Repository name
​
target_name
string
required
Target name
​
template
string
required
Template name: "default" or name of existing target
​
overrides
object
Configuration overrides. See Target Configuration for all fields.
Returns: TargetResponse

​
Batch Create or Update Targets

Create or update multiple targets in a single request. Endpoint: POST /v1/repo/{repo_name}/targets/batch Request: 
curl -X POST "https://bapi.bread.com.ai/v1/repo/my_repo/targets/batch" \
  -H "Authorization: Bearer $BREAD_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "targets": [
        {
            "target_name": "target_alpha",
            "template": "default",
            "overrides": {"model_name": "Qwen/Qwen3-32B"}
        },
        {
            "target_name": "target_beta",
            "template": "target_alpha",
            "overrides": {"temperature": 0.7}
        }
    ]
  }'

​
Create or Update Target (Deprecated)

Deprecated: Use POST /v1/repo/{repo_name}/targets instead. This endpoint will be removed in a future version.
Endpoint: PUT /v1/repo/{repo_name}/targets/{target_name} Request: 
curl -X PUT "https://bapi.bread.com.ai/v1/repo/my_repo/targets/coding_target" \
  -H "Authorization: Bearer $BREAD_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "template": "default",
    "overrides": {
      "model_name": "Qwen/Qwen3-32B",
      "teacher_prompt": "system_prompt_coding",
      "student_prompt": "user_prompt_python"
    }
  }'

​
Batch Create or Update Targets (Deprecated)

Deprecated: Use POST /v1/repo/{repo_name}/targets/batch instead. This endpoint will be removed in a future version.
Endpoint: PUT /v1/repo/{repo_name}/targets/batch Request: 
curl -X PUT "https://bapi.bread.com.ai/v1/repo/my_repo/targets/batch" \
  -H "Authorization: Bearer $BREAD_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "targets": [
      {
        "target_name": "target_alpha",
        "template": "default",
        "overrides": {"model_name": "Qwen/Qwen3-32B"}
      }
    ]
  }'

​
Delete Target

Delete a target from the repository. Endpoint: DELETE /v1/repo/{repo_name}/targets/{target_name} Request: 
curl -X DELETE "https://bapi.bread.com.ai/v1/repo/my_repo/targets/old_target" \
  -H "Authorization: Bearer $BREAD_API_KEY"

​
Configuration Fields

See Target Configuration Reference for detailed documentation of all configuration fields.

​
Key Fields

​
generators
array
Data generation strategies. See Generators.
​
model_name
string
Base model for rollout (required for rollout jobs)
​
teacher_prompt
string
Teacher prompt name - prompt that defines the target behavior being baked into the model.
Deprecated: The u field is deprecated in favor of teacher_prompt. Please use teacher_prompt instead.
​
student_prompt
string
Student prompt name - simplified prompt given at inference. After baking, this triggers the teacher behavior. Can be empty ("") for always-on behavior.
Deprecated: The v field is deprecated in favor of student_prompt. Please use student_prompt instead.
​
max_tokens
integer
Maximum tokens to generate
​
temperature
number
Generation temperature (0.0-2.0)
​
num_traj_per_stimulus
integer
Number of trajectories per stimulus
​
max_concurrency
integer
Maximum concurrent requests

​
Examples

​
Simple Target with Hardcoded Questions

curl -X POST "https://bapi.bread.com.ai/v1/repo/my_repo/targets" \
  -H "Authorization: Bearer $BREAD_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "target_name": "simple_target",
    "template": "default",
    "overrides": {
        "generators": [
            {
                "type": "hardcoded",
                "numq": 3,
                "questions": [
                    "Write a Python function to sort a list",
                    "Create a REST API endpoint",
                    "Implement a binary tree"
                ]
            }
        ],
        "model_name": "Qwen/Qwen3-32B",
      "teacher_prompt": "system_prompt",
      "student_prompt": "user_prompt"
    }
  }'

​
Target with AI-Generated Questions

curl -X POST "https://bapi.bread.com.ai/v1/repo/my_repo/targets" \
  -H "Authorization: Bearer $BREAD_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "target_name": "ai_generated_target",
    "template": "default",
    "overrides": {
        "generators": [
            {
                "type": "oneshot_qs",
                "model": "claude-sonnet-4-5-20250929",
                "numq": 50,
                "temperature": 1.0
            }
        ],
        "model_name": "Qwen/Qwen3-32B",
        "max_tokens": 200,
        "temperature": 0.9,
      "teacher_prompt": "system_coding",
      "student_prompt": "user_task"
    }
  }'
The student prompt can be empty ("") for zero-token, always-on behavior. This proves the expertise is truly baked into the model weights!

​
Target Lifecycle

1

Create Target

Configure target with generators, prompts, and model settings
2

Run Stim

Generate stimuli (questions/tasks) using the target configuration
3

Run Rollout

Generate trajectories (model responses) for the stimuli
4

Use in Bake

Reference the target’s rollout data in a bake for model training

​
Next Steps

Stimulus Generation

Run stim jobs to generate questions

Trajectory Generation

Run rollout jobs for model responses

Generator Types

Configure different generator types

Full Configuration

Complete target configuration reference