MistralRequest

Defined in: adapters/mistral.ts:52

Mistral AI API request structure.

Follows OpenAI-compatible format with some Mistral-specific features like safe_mode and random_seed. Compatible with Mistral’s chat completion endpoint.

See

• MistralMessage
• MistralResponse
• https://docs.mistral.ai/api/

Example

const request: MistralRequest = {
  model: 'mistral-small',
  messages: [
    { role: 'user', content: 'Hello!' }
  ],
  temperature: 0.7,
  max_tokens: 150
};

Properties

max_tokens?

optional max_tokens: number

Defined in: adapters/mistral.ts:63

Maximum tokens to generate

---

messages

messages: MistralMessage[]

Defined in: adapters/mistral.ts:57

Array of conversation messages

---

model

model: string

Defined in: adapters/mistral.ts:54

Model ID (e.g., ‘mistral-small’, ‘mistral-medium’, ‘mistral-large’)

---

random_seed?

optional random_seed: number

Defined in: adapters/mistral.ts:75

Deterministic sampling seed

---

safe_mode?

optional safe_mode: boolean

Defined in: adapters/mistral.ts:72

Enable Mistral’s content safety filter

---

stream?

optional stream: boolean

Defined in: adapters/mistral.ts:69

Enable streaming responses

---

temperature?

optional temperature: number

Defined in: adapters/mistral.ts:60

Sampling temperature 0-1. Higher = more random

---

top_p?

optional top_p: number

Defined in: adapters/mistral.ts:66

Nucleus sampling parameter 0-1