Your first memory in five minutes

Store a belief, query it, and see why it's believed. Every snippet below is real and runs against the same API the dashboard uses.

1. 1

   Install the SDK

   The embedded engine runs offline with full semantics. No key is needed to start.

   PythonTypeScript

   1pip install omem
2. 2

   Create a client

   Point at the cloud with your key, or run fully in-process while prototyping.

   Python

   1import omem

   2 

   3mem = omem.Client() # OMEM_API_KEY from env

   4# mem = omem.Client(embedded=True) # offline, no key
3. 3

   Remember a belief

   Attribute the claim to an agent and, optionally, ground it in an event.

   Python

   1alice = mem.entity("customer:alice")

   2bot = mem.agent("support-bot@v2.1")

   3 

   4mem.remember(bot, about=alice,

   5 claim="prefers_email_over_phone",

   6 because=[mem.event("ticket:8842")])
4. 4

   Query what it believes

   Returns a four-valued state: believed_true, believed_false, contradicted, or unknown.

   Python

   1mem.believes(alice, "prefers_email_over_phone")

   2# -> BELIEVED_TRUE
5. 5

   Ask why

   Get the evidence, the interval, and any contradictions, the same data the dashboard renders.

   Python

   1mem.why(alice, "prefers_email_over_phone")

   2# -> grounded in ticket:8842 / believed since t=1 / no contradictions