mnemosyne-sdk
The official Python client for the Mnemosyne memory substrate. Synchronous and async APIs, container-scoped storage, hybrid BM25+vector recall, full type stubs. Talks to any Mnemosyne endpoint over HTTPS via httpx.
Mnemosyne gives any AI system persistent identity, episodic recall, contradiction resolution, metacognition, and a self-model that compounds across sessions. This SDK is the thin HTTP client that talks to a Mnemosyne Cloud or self-hosted Mnemosyne instance.
Install
pip install mnemosyne-sdk
# or
uv add mnemosyne-sdk
# or
poetry add mnemosyne-sdk
Quick start
from mnemosyne_sdk import Mnemosyne
mn = Mnemosyne(
api_key="mn_...", # from your Mnemosyne dashboard
# base_url="https://api.mnemosyne.geasslabs.xyz", # default
)
# Save a memory
memory = mn.add(
"The dock-A camera lost calibration at 14:32 UTC",
container_tag="fleet-ops",
metadata={"severity": "high"},
)
print(memory.id)
# Search
results = mn.search("when did dock A fail?", container_tag="fleet-ops")
for r in results:
print(f"{r.score:.3f} {r.content}")
# Get / delete
one = mn.get(memory.id)
mn.delete(memory.id)
Async client
import asyncio
from mnemosyne_sdk import AsyncMnemosyne
async def main():
async with AsyncMnemosyne(api_key="mn_...") as mn:
await mn.add("...", container_tag="fleet-ops")
results = await mn.search("query", container_tag="fleet-ops")
asyncio.run(main())
The async client shares the same httpx connection pool and is recommended for high-throughput agents.
Function-calling tools
from mnemosyne_sdk import TOOLS, execute_tool
# Pass TOOLS to any OpenAI-compatible function-calling API
response = client.chat.completions.create(
model="gpt-4",
messages=[...],
tools=TOOLS,
)
# Execute a tool call from the model's response
for call in response.choices[0].message.tool_calls:
result = execute_tool(
call.function.name,
call.function.arguments,
base_url="https://api.mnemosyne.geasslabs.xyz",
api_key="mn_...",
)
Framework adapters
This package is the core client. For framework-specific integrations, install one of the dedicated adapter packages:
| Framework | Package | Install |
|---|---|---|
| LangChain / LangGraph | mnemosyne-langchain |
pip install mnemosyne-sdk[langchain] |
| CrewAI | mnemosyne-crewai |
pip install mnemosyne-sdk[crewai] |
| AutoGen v0.4 | mnemosyne-autogen |
pip install mnemosyne-sdk[autogen] |
| OpenAI Agents SDK | mnemosyne-openai-agents |
pip install mnemosyne-sdk[openai-agents] |
Or install everything: pip install mnemosyne-sdk[all].
Error handling
from mnemosyne_sdk import Mnemosyne, AuthenticationError, RateLimitError
mn = Mnemosyne(api_key="...")
try:
mn.add("...")
except AuthenticationError:
... # bad API key
except RateLimitError as e:
time.sleep(e.retry_after or 1)
... # back off
All errors inherit from MnemosyneError.
Configuration
| Env var | Description | Default |
|---|---|---|
MNEMOSYNE_API_KEY |
API key (overridden by constructor) | - |
MNEMOSYNE_BASE_URL |
Server URL | https://api.mnemosyne.geasslabs.xyz |
MNEMOSYNE_CONTAINER |
Default container tag | default |
Development
git clone https://github.com/synet-systems/mnemosyne
cd mnemosyne/packages/mnemosyne-sdk-python
uv sync --all-extras
uv run pytest
License
MIT © Geass Labs. The Mnemosyne engine this SDK talks to is licensed under the Mnemosyne Source-Available License v1.0. See NOTICE for the dual-license arrangement.
Links
- Homepage: https://mnemosyne.geasslabs.xyz
- Documentation: https://docs.geasslabs.xyz/mnemosyne-sdk
- Repository: https://github.com/synet-systems/mnemosyne
- Issues: https://github.com/synet-systems/mnemosyne/issues
- Changelog: CHANGELOG.md
Metadata
Release files for mnemosyne-sdk 0.1.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| mnemosyne_sdk-0.1.1.tar.gz | 7.2 kB | Without provenance | Details for mnemosyne_sdk-0.1.1.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| mnemosyne_sdk-0.1.1-py3-none-any.whl | Python 3 | none | any | Details for mnemosyne_sdk-0.1.1-py3-none-any.whl |
Total release size: 14.6 kB
Release files / mnemosyne_sdk-0.1.1.tar.gz
| Download URL | mnemosyne_sdk-0.1.1.tar.gz |
|---|---|
| Size | 7.2 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
805b3e232edf827e8b1bae6e65e4221ade1c482dd2701e629b570a33ef72e183
|
|
BLAKE2b-256 checksum How to use checksums |
6b84abdaa3164a331fe7df9d3bd1e38bc71cfe06774ac1c12aa86647bf8422dc
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.14.2
|
Release files / mnemosyne_sdk-0.1.1-py3-none-any.whl
| Download URL | mnemosyne_sdk-0.1.1-py3-none-any.whl |
|---|---|
| Size | 7.4 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
5049875ace2b643a607fbfb1eb1b719f23c6af963d19b9ec0ec278d7acf0ced4
|
|
BLAKE2b-256 checksum How to use checksums |
03407b0028a9e7cf924b3f2f2ae1dc00cfde5943923bba634fed8fcf0966c77e
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.2.0 CPython/3.14.2
|
Security for mnemosyne-sdk 0.1.1
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.