Quickstart: Add AI Memory in Minutes

Get the Aletheia engine running, connect your application using our SDK, and start giving your AI agents persistent, time-aware memory.

The Local-First Development Path

We believe developers should be able to build and test without relying on external cloud services. Aletheia is built in Rust, meaning it compiles to a blazing-fast, single binary you can run right on your machine.

First, build and run the engine locally:

cargo build --release
cargo run --release

Next, point our Python SDK at your local instance. In just three lines of code, you can ingest a new fact and retrieve it:

from aletheia import AletheiaClient

# Connect to the local Aletheia binary
client = AletheiaClient.from_local(auto_start=True)

# Give your agent a permanent memory
client.ingest(entity_id="user-123", text="I prefer pourover coffee.")

# Retrieve the exact truth when you need it
hits = client.query("What coffee do I prefer?", entity_id="user-123")

Seamless Cloud Deployment

When you are ready to ship to production, you don't need to rewrite your memory logic. Simply create an API key in the Aletheia platform dashboard, and switch your SDK constructor to point to the cloud.

client = AletheiaClient.from_cloud(
    "https://api.aletheia.com", # Your hosted endpoint
    api_key="YOUR_PRODUCTION_API_KEY",
)

Our goal is zero friction: the developer API surface remains completely stable whether you are running a test script on your laptop or serving millions of users in production.

Ready to dive deeper? Check out our guides on Fact Supersession and Temporal Retrieval to see what makes Aletheia smarter than a standard vector database.