Plugs into your stack.

from openai import OpenAI
from feather_db import DB
import numpy as np

client = OpenAI()
db = DB.open("embeddings.feather", dim=1536)

resp = client.embeddings.create(
    model="text-embedding-3-small",
    input="Your text here",
)
vec = np.array(resp.data[0].embedding, dtype=np.float32)

db.add(id=1, vec=vec)
db.save()

# Embed with voyage-3, retrieve with Feather, answer with Claude
from voyageai import Client as Voyage
from anthropic import Anthropic
from feather_db import DB

voyage = Voyage()
claude = Anthropic()
db = DB.open("rag.feather", dim=1024)

def ingest(text: str, id: int):
    e = voyage.embed([text], model="voyage-3").embeddings[0]
    db.add(id=id, vec=e)

ctx = db.context_chain(query_vec, k=5, hops=2)
msg = claude.messages.create(model="claude-sonnet-4-6", ...)

import google.generativeai as genai
from feather_db import DB

db = DB.open("embeddings.feather", dim=768)

r = genai.embed_content(
    model="gemini-embedding-exp-03-07",
    content="Your text here",
)
db.add(id=1, vec=r["embedding"])

from sentence_transformers import SentenceTransformer
from feather_db import DB
import numpy as np

model = SentenceTransformer("all-MiniLM-L6-v2")
db = DB.open("embeddings.feather", dim=384)

text = "Your text here"
vec = model.encode(text).astype(np.float32)
db.add(id=1, vec=vec)

import ollama
from feather_db import DB
import numpy as np

db = DB.open("embeddings.feather", dim=1024)
r = ollama.embeddings(model="nomic-embed-text", prompt="Your text")
vec = np.array(r["embedding"], dtype=np.float32)
db.add(id=1, vec=vec)

from langchain.vectorstores import FeatherVectorStore
from langchain.embeddings import OpenAIEmbeddings

emb = OpenAIEmbeddings()
store = FeatherVectorStore(embedding=emb, db_path="langchain.feather")

store.add_texts(["Document 1", "Document 2"])
results = store.similarity_search("query", k=5)

from langgraph.graph import StateGraph
from feather_db import DB

db = DB.open("agent_memory.feather", dim=768)

def retrieve(state):
    ctx = db.context_chain(state["query_vec"], k=5, hops=2)
    return {"context": ctx}

graph = StateGraph(state_schema).add_node("retrieve", retrieve)

from crewai import Crew, Agent
from feather_db import DB

db = DB.open("crew_memory.feather", dim=768)

researcher = Agent(
    role="researcher",
    memory=lambda q: db.context_chain(q, k=5, hops=2),
)
crew = Crew(agents=[researcher], verbose=True)

from llama_index.vector_stores import FeatherVectorStore
from llama_index import VectorStoreIndex

store = FeatherVectorStore(db_path="llamaindex.feather")
index = VectorStoreIndex.from_vector_store(store)
engine = index.as_query_engine()
print(engine.query("Your question"))

import { DB } from "feather-db";

const db = await DB.open("vectors.feather", { dim: 384 });
await db.add(1, new Float32Array(384));
const results = await db.search(query, { k: 5 });

import init, { DB } from "feather-wasm";

await init();
const db = new DB("vectors.feather", 384);
db.add(1, new Float32Array(384));
const results = db.search(query, 5);

// Swift (iOS)
import Feather

let db = try DB.open(path: "vectors.feather", dim: 384)
try db.add(id: 1, vector: vector)
try db.save()