Langbase

Langbase is a serverless AI developer platform for building, deploying, and scaling AI agents and applications. Its composable primitives - Pipes (agents), Memory (managed RAG), Threads, Agent (one API over 100+ LLMs), Tools, Parser, Chunker, and Embed - are exposed through a single Bearer-authenticated REST API at api.langbase.com, with Server-Sent Events (SSE) streaming for generative endpoints.

9 APIs 0 Features

AIAgentsRAGLLMServerless

APIs

Langbase Pipes API

Run, create, update, and list Pipes - composable, deployable AI agents with a system prompt, model, tools, variables, and memory. POST /v1/pipes/run executes a Pipe over a messa...

Langbase Agent API

A single liquid AI mesh endpoint (POST /v1/agent/run) that runs a prompt against 100+ LLMs from all top providers using a provider:model_id string, with tools, generation contro...

Langbase Memory API

Managed, serverless RAG. Create and list memory stores, upload and manage documents, and retrieve semantically similar chunks (POST /v1/memory/retrieve) with a topK control and ...

Langbase Threads API

Persist and manage conversation state. Create, get, update, and delete threads, and append or list messages, so multi-turn agent conversations keep their history server-side and...

Langbase Tools API

Managed agent tools. POST /v1/tools/web-search runs a live web search (via providers such as Exa) returning URLs and extracted content, and POST /v1/tools/crawl fetches and extr...

Langbase Parser API

Extract clean text from uploaded documents (PDF, CSV, XLSX, XLS, and more, up to 10 MB) via a multipart POST /v1/parser, as the ingestion front door for RAG and content pipelines.

Langbase Chunker API

Split large text into overlapping, retrieval-ready chunks (POST /v1/chunker) with configurable chunkMaxLength and chunkOverlap, for building RAG and search indexes.

Langbase Embed API

Generate embedding vectors for up to 100 text chunks per request (POST /v1/embed) using OpenAI, Cohere, or Google embedding models, defaulting to openai:text-embedding-3-large.

Langbase Workflows API

Durable, step-based orchestration for composing agents, memory, and tools into multi-step AI workflows with retries and timeouts. Workflows compose the underlying Agent, Pipes, ...

Collections

Langbase API

OPEN

Pricing Plans

Langbase Plans Pricing

4 plans

PLANS

Rate Limits

Langbase Rate Limits

6 limits

RATE LIMITS

FinOps

Langbase Finops

FINOPS

Resources

🔗

VulnerabilityDisclosure

VulnerabilityDisclosure

GitHubOrganization

Sources

opencollection: 1.0.0
info:
  name: Langbase API
  version: '1.0'
request:
  auth:
    type: bearer
    token: '{{bearerToken}}'
items:
- info:
    name: Pipes
    type: folder
  items:
  - info:
      name: Run a Pipe.
      type: http
    http:
      method: POST
      url: https://api.langbase.com/v1/pipes/run
      body:
        type: json
        data: "{\n  \"messages\": [{ \"role\": \"user\", \"content\": \"Hello\" }],\n  \"stream\": false\n}"
    docs: Executes a Pipe over a message array. Streams Server-Sent Events when stream is true.
  - info:
      name: List Pipes.
      type: http
    http:
      method: GET
      url: https://api.langbase.com/v1/pipes
    docs: List Pipes.
  - info:
      name: Create a Pipe.
      type: http
    http:
      method: POST
      url: https://api.langbase.com/v1/pipes
      body:
        type: json
        data: "{\n  \"name\": \"my-pipe\"\n}"
    docs: Create a new Pipe by sending the Pipe configuration in the request body.
  - info:
      name: Update a Pipe.
      type: http
    http:
      method: POST
      url: https://api.langbase.com/v1/pipes/:name
      params:
      - name: name
        value: ''
        type: path
        description: The name of the Pipe to update.
      body:
        type: json
        data: '{}'
    docs: Update an existing Pipe.
- info:
    name: Agent
    type: folder
  items:
  - info:
      name: Run an agent against any of 100+ LLMs.
      type: http
    http:
      method: POST
      url: https://api.langbase.com/v1/agent/run
      body:
        type: json
        data: "{\n  \"model\": \"openai:gpt-4o-mini\",\n  \"input\": \"Hello\",\n  \"stream\": false\n}"
    docs: Runs a prompt against a provider:model_id. Streams Server-Sent Events when stream is true. Supports LB-LLM-Key for
      bring-your-own-key.
- info:
    name: Memory
    type: folder
  items:
  - info:
      name: List memory stores.
      type: http
    http:
      method: GET
      url: https://api.langbase.com/v1/memory
    docs: List memory stores.
  - info:
      name: Create a memory store.
      type: http
    http:
      method: POST
      url: https://api.langbase.com/v1/memory
      body:
        type: json
        data: "{\n  \"name\": \"my-memory\"\n}"
    docs: Create a memory store.
  - info:
      name: Retrieve similar chunks (RAG).
      type: http
    http:
      method: POST
      url: https://api.langbase.com/v1/memory/retrieve
      body:
        type: json
        data: "{\n  \"query\": \"What is Langbase?\",\n  \"memory\": [{ \"name\": \"my-memory\" }],\n  \"topK\": 20\n}"
    docs: Retrieve semantically similar chunks from one or more memory stores.
  - info:
      name: Delete a memory store.
      type: http
    http:
      method: DELETE
      url: https://api.langbase.com/v1/memory/:name
      params:
      - name: name
        value: ''
        type: path
    docs: Delete a memory store.
  - info:
      name: List documents in a memory store.
      type: http
    http:
      method: GET
      url: https://api.langbase.com/v1/memory/:name/documents
      params:
      - name: name
        value: ''
        type: path
    docs: List documents in a memory store.
  - info:
      name: Upload a document to a memory store.
      type: http
    http:
      method: POST
      url: https://api.langbase.com/v1/memory/:name/documents
      params:
      - name: name
        value: ''
        type: path
      body:
        type: multipart-form
        data: []
    docs: Upload a document to a memory store.
  - info:
      name: Delete a document from a memory store.
      type: http
    http:
      method: DELETE
      url: https://api.langbase.com/v1/memory/:name/documents/:documentName
      params:
      - name: name
        value: ''
        type: path
      - name: documentName
        value: ''
        type: path
    docs: Delete a document from a memory store.
- info:
    name: Threads
    type: folder
  items:
  - info:
      name: Create a conversation thread.
      type: http
    http:
      method: POST
      url: https://api.langbase.com/v1/threads
      body:
        type: json
        data: '{}'
    docs: Create a conversation thread.
  - info:
      name: Get a thread.
      type: http
    http:
      method: GET
      url: https://api.langbase.com/v1/threads/:threadId
      params:
      - name: threadId
        value: ''
        type: path
    docs: Get a thread.
  - info:
      name: Update a thread.
      type: http
    http:
      method: PATCH
      url: https://api.langbase.com/v1/threads/:threadId
      params:
      - name: threadId
        value: ''
        type: path
      body:
        type: json
        data: '{}'
    docs: Update a thread.
  - info:
      name: Delete a thread.
      type: http
    http:
      method: DELETE
      url: https://api.langbase.com/v1/threads/:threadId
      params:
      - name: threadId
        value: ''
        type: path
    docs: Delete a thread.
  - info:
      name: List messages in a thread.
      type: http
    http:
      method: GET
      url: https://api.langbase.com/v1/threads/:threadId/messages
      params:
      - name: threadId
        value: ''
        type: path
    docs: List messages in a thread.
  - info:
      name: Append messages to a thread.
      type: http
    http:
      method: POST
      url: https://api.langbase.com/v1/threads/:threadId/messages
      params:
      - name: threadId
        value: ''
        type: path
      body:
        type: json
        data: '[]'
    docs: Append messages to a thread.
- info:
    name: Tools
    type: folder
  items:
  - info:
      name: Run a live web search.
      type: http
    http:
      method: POST
      url: https://api.langbase.com/v1/tools/web-search
      body:
        type: json
        data: "{\n  \"query\": \"Langbase\",\n  \"service\": \"exa\"\n}"
    docs: Run a live web search, returning URLs and extracted content.
  - info:
      name: Crawl and extract content from web pages.
      type: http
    http:
      method: POST
      url: https://api.langbase.com/v1/tools/crawl
      body:
        type: json
        data: "{\n  \"url\": [\"https://langbase.com\"],\n  \"service\": \"spider\"\n}"
    docs: Crawl and extract content from one or more web pages.
- info:
    name: Parser
    type: folder
  items:
  - info:
      name: Parse a document into clean text.
      type: http
    http:
      method: POST
      url: https://api.langbase.com/v1/parser
      body:
        type: multipart-form
        data: []
    docs: Extract clean text from an uploaded document (PDF, CSV, XLSX, XLS, and more, up to 10 MB).
- info:
    name: Chunker
    type: folder
  items:
  - info:
      name: Split text into overlapping chunks.
      type: http
    http:
      method: POST
      url: https://api.langbase.com/v1/chunker
      body:
        type: json
        data: "{\n  \"content\": \"...\",\n  \"chunkMaxLength\": 1024,\n  \"chunkOverlap\": 256\n}"
    docs: Split large text into overlapping, retrieval-ready chunks.
- info:
    name: Embed
    type: folder
  items:
  - info:
      name: Generate embedding vectors.
      type: http
    http:
      method: POST
      url: https://api.langbase.com/v1/embed
      body:
        type: json
        data: "{\n  \"chunks\": [\"text to embed\"],\n  \"embeddingModel\": \"openai:text-embedding-3-large\"\n}"
    docs: Generate embedding vectors for up to 100 chunks per request.
bundled: true