AI & Text Generation

Tether exposes AI helpers for drafting replies and testing prompts. Production traffic runs through your organization's configured model and credentials, while the playground lets you iterate on prompts before wiring them into workflows.

POST/api/ai/process

Generate a reply for a conversation

Generates an AI reply based on conversation history, the assigned process prompt, and optional additional context. Authorization checks ensure only owners, assignees, admins, or superadmins can generate responses.

Required attributes

Name
conversationId
Type
string
Description
Conversation to generate a reply for.

Optional attributes

Name
mergeContextPrompt
Type
string
Description
Extra context prepended to the process prompt.

Request

POST

/api/ai/process

curl -X POST https://your-tether-instance.com/api/ai/process \
  -H "Authorization: Bearer {token}" \
  -H "Content-Type: application/json" \
  -d '{
    "conversationId": "6611152b68d5160f12345678",
    "mergeContextPrompt": "Prioritize booking a test drive this week."
  }'

POST/api/ai/playground

Run a playground prompt

Send ad-hoc messages to the AI using any saved prompt. Optionally include a lightweight conversation array to preserve context between turns.

Required attributes

Name
message
Type
string
Description
User message to send to the model.
Name
promptId
Type
string
Description
Prompt to apply when generating.

Optional attributes

Name
conversation
Type
array
Description
Array of { role, content } entries to include as prior turns.

Request

POST

/api/ai/playground

curl -X POST https://your-tether-instance.com/api/ai/playground \
  -H "Authorization: Bearer {token}" \
  -H "Content-Type: application/json" \
  -d '{
    "message": "Write a follow-up SMS confirming tomorrow's appointment",
    "promptId": "66111473a6e3b40f12345678",
    "conversation": [
      { "role": "user", "content": "Lead booked a test drive for Friday 2pm." }
    ]
  }'