Skip to main content

Long-term memory for AI agents. v2.3 adds Inheritance + Federation, MCP v2.3 (10 new tools), L2 Inject endpoint, Vietnamese FTS.

Project description

MemoryAI — Long-term Memory for AI Agents

Your AI never forgets. Preferences, decisions, context — remembered forever.

Install

Note: Package name is hmc-memory because memoryai was squatted on PyPI in 2024.
Import name remains memoryai for clean code: from memoryai import MemoryAI.

pip install hmc-memory

Quick Start

from memoryai import MemoryAI

mem = MemoryAI(api_key="hm_sk_your_key")

# Store a memory
mem.store("User prefers dark mode", memory_type="preference")

# Recall
results = mem.recall("what does user prefer")
for r in results:
    print(f"[{r['score']:.0%}] {r['content']}")

# Bootstrap (start of session)
context = mem.bootstrap("working on payment feature")
# → Returns ~800 token context block with identity + mood + goals

# Save (end of session)
mem.save("Built payment API with Stripe. User decided on webhook approach.")

# Profile
profile = mem.profile()
print(profile["persona"])
print(profile["mood"])

API

MemoryAI(api_key, endpoint, timeout, graceful)

Param Default Description
api_key required Your API key (hm_sk_...)
endpoint https://memoryai.dev API endpoint
timeout 30 Request timeout (seconds)
graceful False Never raise errors, return empty instead

Methods

Method Description
store(content, memory_type, tags) Store a memory
recall(query, limit, depth, since) Search memories
bootstrap(task, mode) Wake up with context (session start)
save(content) Save session summary (session end)
profile() Get cognitive profile (persona, mood, goals)
health() Check memory stats
guard_check(tokens, max_tokens) Check context pressure

Memory Types

Type Lifespan Use for
preference Forever User likes/dislikes
decision Forever Choices made
identity Forever Who the user is
procedure Forever How to do things
fact Decays General knowledge
goal Until done Active objectives

Zero Dependencies

Pure Python stdlib (urllib). No requests, no httpx, no aiohttp. Works everywhere Python runs.

Links

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

hmc_memory-2.3.0.tar.gz (7.0 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

hmc_memory-2.3.0-py3-none-any.whl (7.5 kB view details)

Uploaded Python 3

File details

Details for the file hmc_memory-2.3.0.tar.gz.

File metadata

  • Download URL: hmc_memory-2.3.0.tar.gz
  • Upload date:
  • Size: 7.0 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.3

File hashes

Hashes for hmc_memory-2.3.0.tar.gz
Algorithm Hash digest
SHA256 4d14bfc179be2929bbab5109cd2a9dcbbfba3360c2751345b814a557d8acb1de
MD5 1eb4a21842ed0a2cb935fe5d0a5c482e
BLAKE2b-256 f6ca60ccd72962b6cca00c43e4d0232a03fcb7a3a3e430b43e6e15b0b9416baf

See more details on using hashes here.

File details

Details for the file hmc_memory-2.3.0-py3-none-any.whl.

File metadata

  • Download URL: hmc_memory-2.3.0-py3-none-any.whl
  • Upload date:
  • Size: 7.5 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.3

File hashes

Hashes for hmc_memory-2.3.0-py3-none-any.whl
Algorithm Hash digest
SHA256 70a73ce5a349476f5fa1305e0498987b0096c3e78f9149c89109e082b5b09189
MD5 9841795f6fa70d01fa3236fba11fbbff
BLAKE2b-256 ba9734be4a88b4f6deb740f6a47e5c2ff67b25db16e39eccd935d0d7cf0efe79

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page