SDK Overview

Flow-Like provides official SDKs for Node.js/TypeScript and Python that wrap the REST API into ergonomic, type-safe clients. Use them to control workflows, manage files, query LanceDB databases, run chat completions, generate embeddings, and more — all from your application code.

Available SDKs

Language	Package	Install
Node.js / TypeScript	`@flow-like/sdk`	`npm install @flow-like/sdk`
Python	`flow-like`	`uv add flow-like`

Feature Matrix

Feature	Node.js	Python
Trigger workflows (sync SSE + async)	✅	✅
Trigger events	✅	✅
File management (upload, download, list, delete)	✅	✅
LanceDB integration (credentials, queries, connections)	✅	✅
Chat completions (+ streaming)	✅	✅
Embeddings	✅	✅
Model discovery (LLMs, embeddings)	✅	✅
Board CRUD	✅	✅
App management	✅	✅
HTTP sinks	✅	✅
Execution monitoring / polling	✅	✅
LangChain integration	✅	✅
Async / await	✅ native	✅ `a`-prefixed methods

Authentication

Both SDKs support two authentication methods:

Type	Prefix	Header sent	Env variable
Personal Access Token	`pat_`	`Authorization: pat_{id}.{secret}`	`FLOW_LIKE_PAT`
API Key	`flk_`	`X-API-Key: flk_{app}.{key}.{secret}`	`FLOW_LIKE_API_KEY`

The SDK auto-detects which header to use based on the token prefix.

Environment variables

Both SDKs read from environment variables by default:

export FLOW_LIKE_BASE_URL=https://api.flow-like.com
export FLOW_LIKE_PAT=pat_myid.mysecret
# or
export FLOW_LIKE_API_KEY=flk_appid.keyid.secret

LangChain Integration

Both SDKs include optional LangChain-compatible wrappers so you can use Flow-Like models inside LangChain chains, agents, and RAG pipelines.

Node.js: import { FlowLikeChatModel, FlowLikeEmbeddings } from "@flow-like/sdk/langchain"
Python: from flow_like.langchain import FlowLikeChatModel, FlowLikeEmbeddings

See the language-specific pages for installation and usage details.