POST
/
v1
/
prompt_engine
/
run
Run Prompt Direct is ideal for one-time prompt usage.

For prompts that will be used multiple times, it is recommended to first Create a Prompt and then Run the Prompt for better reliability.

​
x-api-key
string
required

Your JigsawStack API key

​
Body

​
prompt
string
required

The prompt. Maximum character limit is 3000. Prompt supports dynamic inputs. See example below.

​
inputs
array<object>

The prompt inputs. See example below.

​
return_prompt
string | array | object

How the prompt result should be returned or formatted. See examples below

​
prompt_guard
array<string>

Include this to guard against unsafe inputs from users. Supported values:

  • defamation
  • privacy
  • hate
  • sexual_content
  • elections
  • code_interpreter_abuse
  • indiscrimate_weapons
  • specialized_advice
​
input_values
object

Key-value pair for dynamic prompt variables.

​
stream
boolean

If set, partial message chunk will be sent.

​
use_internet
boolean

Include this to allow prompt engine to use the internet.

​
Sample Prompt Payload

  • String return_prompt
{
    prompt: "Tell me a story about {about}",
    inputs: [
      {
        key: "about",
        optional: false,
      },
    ],
    return_prompt: "Return the result in a markdown format",
    "prompt_guard": ["sexual_content", "defamation"]
    "input_values": {
      "about": "Santorini",
    }
  }
  • Array<object> return_prompt

{
    prompt: "Tell me a story about {about}",
    inputs: [
      {
        key: "about",
        optional: false,
        initial_value: "Leaning Tower of Pisa",
      },
    ],
    return_prompt: [{ excerpt: "short story text", summary: "summary of story" }],
    "prompt_guard": ["sexual_content", "defamation"],
    "input_values": {
      "about": "Santorini",
    }
  }
  • Object return_prompt
{
    prompt: "Tell me a story about {about}",
    inputs: [
      {
        key: "about",
        optional: false,
        initial_value: "Leaning Tower of Pisa",
      },
    ],
    return_prompt: { excerpt: "short story text", summary: "summary of story" },
    "prompt_guard": ["sexual_content", "defamation"],
    "input_values": {
      "about": "Santorini",
    }
  }

​
Response

​
success
boolean

Indicates whether the call was successful.

Was this page helpful?

Delete a PromptvOCR