Lamini

Lamini is an enterprise LLM platform for fine-tuning, tuning, and serving custom open models. Its Memory Tuning approach embeds factual recall into models to reduce hallucination, and the platform exposes a REST API for inference (completions), fine-tuning jobs, classification, and embeddings over open base models, deployable in Lamini's cloud, on-demand GPU cluster, or on-premises.

4 APIs 0 Features

AILLMFine-TuningMemory TuningInference

APIs

Lamini Inference Completions API

Generate text completions from open base or tuned models via POST /v1/completions, with structured (typed) output via output_type, plus streaming completions at /v3/streaming_co...

Lamini Fine-Tuning & Memory Tuning API

Submit and manage tuning jobs against open base models via POST /v1/train, supporting full fine-tuning and Lamini Memory Tuning (train_type), with job listing, status, cancel, a...

Lamini Classify API

Run LLM-based text classification against a trained classifier model via POST /v1/classifier/{model_id}/classification for scored labels, or /v1/classifier/{model_id}/prediction...

Lamini Embeddings API

Encode one or more text prompts into embedding vectors via POST /v1/embedding for similarity search, retrieval, and Memory RAG indexing workflows.

Collections

Lamini Platform API

OPEN

Pricing Plans

Lamini Plans Pricing

2 plans

PLANS

Rate Limits

Lamini Rate Limits

4 limits

RATE LIMITS

FinOps

Lamini Finops

FINOPS

Resources

👥

GitHubOrganization

Sources

opencollection: 1.0.0
info:
  name: Lamini Platform API
  version: '1.0'
request:
  auth:
    type: bearer
    token: '{{bearerToken}}'
items:
- info:
    name: Inference
    type: folder
  items:
  - info:
      name: Generate a completion
      type: http
    http:
      method: POST
      url: https://api.lamini.ai/v1/completions
      body:
        type: json
        data: "{\n  \"prompt\": \"What is the capital of France?\",\n  \"model_name\": \"meta-llama/Llama-3.1-8B-Instruct\"\
          ,\n  \"output_type\": null,\n  \"max_new_tokens\": 128\n}"
    docs: Generate a text completion from a base or tuned model. Supports a typed output schema via output_type for structured
      generation.
  - info:
      name: Generate a streaming completion
      type: http
    http:
      method: POST
      url: https://api.lamini.ai/v3/streaming_completions
      body:
        type: json
        data: "{\n  \"prompt\": \"Write a short poem about APIs.\",\n  \"model_name\": \"meta-llama/Llama-3.1-8B-Instruct\"\
          ,\n  \"max_new_tokens\": 256\n}"
    docs: Generate a completion as an incremental stream of token chunks for the provided prompt and model.
- info:
    name: Fine-Tuning
    type: folder
  items:
  - info:
      name: Submit a tuning job
      type: http
    http:
      method: POST
      url: https://api.lamini.ai/v1/train
      body:
        type: json
        data: "{\n  \"model_name\": \"meta-llama/Llama-3.1-8B-Instruct\",\n  \"dataset_id\": \"your-dataset-id\",\n  \"train_type\"\
          : \"memory_rag\"\n}"
    docs: Submit a fine-tuning or Memory Tuning job for a base model against an uploaded dataset. The train_type field selects
      the tuning method.
  - info:
      name: List tuning jobs
      type: http
    http:
      method: GET
      url: https://api.lamini.ai/v1/train/jobs
    docs: List tuning jobs for the authenticated account.
  - info:
      name: Get a tuning job
      type: http
    http:
      method: GET
      url: https://api.lamini.ai/v1/train/jobs/{job_id}
    docs: Retrieve the status and metadata of a single tuning job.
  - info:
      name: Cancel a tuning job
      type: http
    http:
      method: POST
      url: https://api.lamini.ai/v1/train/jobs/{job_id}/cancel
    docs: Cancel a running tuning job by its job id.
  - info:
      name: Resume a tuning job
      type: http
    http:
      method: POST
      url: https://api.lamini.ai/v1/train/jobs/{job_id}/resume
    docs: Resume a previously paused or cancelled tuning job.
- info:
    name: Classify
    type: folder
  items:
  - info:
      name: Classify text
      type: http
    http:
      method: POST
      url: https://api.lamini.ai/v1/classifier/{model_id}/classification
      body:
        type: json
        data: "{\n  \"prompt\": \"My order never arrived and I want a refund.\",\n  \"top_n\": 3\n}"
    docs: Classify one or more text prompts with a trained classifier model, returning scored class labels.
  - info:
      name: Predict a class
      type: http
    http:
      method: POST
      url: https://api.lamini.ai/v1/classifier/{model_id}/prediction
      body:
        type: json
        data: "{\n  \"prompt\": \"My order never arrived and I want a refund.\"\n}"
    docs: Return the single predicted class for one or more text prompts using a trained classifier model.
- info:
    name: Embeddings
    type: folder
  items:
  - info:
      name: Generate embeddings
      type: http
    http:
      method: POST
      url: https://api.lamini.ai/v1/embedding
      body:
        type: json
        data: "{\n  \"prompt\": [\"The quick brown fox.\", \"Lorem ipsum dolor sit amet.\"]\n}"
    docs: Encode one or more text prompts into embedding vectors for similarity search, retrieval, and indexing.
- info:
    name: Platform
    type: folder
  items:
  - info:
      name: Get platform version
      type: http
    http:
      method: GET
      url: https://api.lamini.ai/v1/version
    docs: Return server and client version metadata and the list of available API endpoints and their versions.