Not Diamond

Not Diamond is an AI model router that determines the best LLM to call for any given prompt. Its REST API routes each request to the optimal model across providers based on quality, cost, and latency tradeoffs, accepts real-time feedback to personalize routing, and can train custom routers from evaluation datasets.

4 APIs 0 Features

AILLMModel RoutingRouterOrchestration

APIs

Not Diamond Model Select API

Routes a prompt to the best available LLM. Submit OpenAI-format messages and a list of candidate providers, and the modelSelect endpoint returns the recommended provider/model p...

Not Diamond Models API

Lists supported text generation models with provider, context length, and per-million-token input/output pricing, with optional filtering by provider or OpenRouter support.

Not Diamond Feedback API

Reports outcome feedback and latency metrics against a routing session so Not Diamond can continuously personalize routing decisions in real time to an application's end users.

Not Diamond Custom Routers API

Trains a custom router from an uploaded CSV evaluation dataset and creates or updates a user preference, returning a preference ID used to drive personalized routing in subseque...

Collections

Not Diamond API

OPEN

Pricing Plans

Notdiamond Plans Pricing

2 plans

PLANS

Rate Limits

Notdiamond Rate Limits

3 limits

RATE LIMITS

FinOps

Notdiamond Finops

FINOPS

Resources

👥

GitHubOrganization

Sources

opencollection: 1.0.0
info:
  name: Not Diamond API
  version: '2.0'
request:
  auth:
    type: bearer
    token: '{{bearerToken}}'
items:
- info:
    name: Model Routing
    type: folder
  items:
  - info:
      name: Route a prompt to the best model.
      type: http
    http:
      method: POST
      url: https://api.notdiamond.ai/v2/modelRouter/modelSelect
      body:
        type: json
        data: "{\n  \"messages\": [\n    { \"role\": \"user\", \"content\": \"<string>\" }\n  ],\n  \"llm_providers\": [\n\
          \    { \"provider\": \"openai\", \"model\": \"gpt-4o\" },\n    { \"provider\": \"anthropic\", \"model\": \"claude-3-5-sonnet-20241022\"\
          \ }\n  ],\n  \"tradeoff\": \"cost\",\n  \"max_model_depth\": 2,\n  \"hash_content\": false,\n  \"metric\": \"accuracy\"\
          \n}"
    docs: Given OpenAI-format messages and a list of candidate LLM providers, returns the recommended provider/model and a
      session ID.
- info:
    name: Models
    type: folder
  items:
  - info:
      name: List supported models.
      type: http
    http:
      method: GET
      url: https://api.notdiamond.ai/v2/models
    docs: Lists supported text generation models with provider, context length, and per-million-token pricing.
- info:
    name: Feedback
    type: folder
  items:
  - info:
      name: Report outcome feedback for a routing session.
      type: http
    http:
      method: POST
      url: https://api.notdiamond.ai/v2/report/metrics/feedback
      body:
        type: json
        data: "{\n  \"session_id\": \"<string>\",\n  \"provider\": { \"provider\": \"openai\", \"model\": \"gpt-4o\" },\n\
          \  \"feedback\": { \"value\": 1 }\n}"
    docs: Submits binary outcome feedback (0 or 1) for the provider selected in a routing session.
  - info:
      name: Report latency metrics for a routing session.
      type: http
    http:
      method: POST
      url: https://api.notdiamond.ai/v2/report/metrics/latency
      body:
        type: json
        data: "{\n  \"session_id\": \"<string>\",\n  \"provider\": { \"provider\": \"openai\", \"model\": \"gpt-4o\" },\n\
          \  \"feedback\": { \"tokens_per_second\": 0 }\n}"
    docs: Reports observed latency (e.g. tokens per second) for the provider selected in a routing session.
- info:
    name: Custom Routers
    type: folder
  items:
  - info:
      name: Train a custom router.
      type: http
    http:
      method: POST
      url: https://api.notdiamond.ai/v2/pzn/trainCustomRouter
      body:
        type: multipart-form
        data: []
    docs: Trains a custom router from an uploaded CSV evaluation dataset and returns a preference ID.
  - info:
      name: Create a user preference.
      type: http
    http:
      method: POST
      url: https://api.notdiamond.ai/v2/preferences/userPreferenceCreate
      body:
        type: json
        data: "{\n  \"name\": \"<string>\"\n}"
    docs: Creates a user preference (router) and returns its preference ID.