engram-client

Python client for Engram Cloud - AI memory infrastructure for agents.

Installation

pip install engram-client

Quick Start

from engram_client import EngramClient

client = EngramClient(
    base_url="https://engram-cloud-gateway.fly.dev",
    api_key="ek_...",
    tenant="my-tenant",
)

# Create a memory
memory = client.create(
    "User prefers dark mode",
    tags=["preferences", "ui"],
    workspace="my-project",
)

# Search (hybrid: BM25 + vector + fuzzy)
results = client.search("user preferences")

# List with filters
memories = client.list(limit=20, workspace="my-project")

# Get by ID
memory = client.get(42)

# Update
client.update(42, content="User prefers light mode", tags=["preferences"])

# Delete
client.delete(42)

# Stats
stats = client.stats()

Context Manager

with EngramClient(base_url="...", api_key="...", tenant="...") as client:
    client.create("Hello from Python SDK")

API Reference

EngramClient(base_url, api_key, tenant)

Method	Description
create(content, **kwargs)	Create a memory
get(id)	Get memory by ID
update(id, **kwargs)	Update a memory
delete(id)	Delete a memory
list(**kwargs)	List memories with filters
search(query, **kwargs)	Hybrid search
stats()	Storage statistics

Parameters

create / update kwargs: tags, workspace, memory_type, importance, metadata, tier

list kwargs: limit, offset, workspace, memory_type, tags, sort_by, sort_order

search kwargs: limit, workspace, tags, memory_type, include_archived

Requirements

• Python >= 3.9
• httpx >= 0.25.0

Related

• Engram - Core memory engine (Rust)
• Engram Cloud - Multi-tenant SaaS gateway
• engram-client - TypeScript client

License

MIT