letta-client

library1.12.1pypypi✓ verified 80d ago

The official Python library for the Letta API, providing a typed interface to agents, conversations, messages, tools, and more. Current version 1.10.3, requires Python >=3.9, actively maintained with frequent releases (multiple per month).

llm-agents ai-ml

pip install letta-client

github ↗package ↗homepage ↗

INSTALL

IMPORT

SIG · LETTA-CLIENT

letta-client

llm-agentspythonv1.12.1

Install

4.2s avg

Import

—

Disk

35MB

Pass rate

10/ 10

Env Coverage10 / 10

glibc

3.9–3.13

musl

3.9–3.13

Install & Compatibility

Where this runs

tested against v1.12.1 · pip install

no network on importno background threads

Install × environment matrix

Each cell = how many times install + import succeeded across repeated harness runs. Partial = flaky.
glibc = Debian/Ubuntu slim · musl = Alpine Linux

musl

py 3.10–3.95 runs

installs and imports cleanly · install 0.0s · import 0.000s · 36.8MB

glibc

py 3.10–3.95 runs

installs and imports cleanly · install 4.2s · import 0.000s · 36MB

35MB installed

● package 35MB

Code

Verified usage

Verified import paths — ran on the pinned version, not inferred.

Client

✓ from letta_client import Client

✗ from letta import Letta

AsyncClient

✓ from letta_client import AsyncClient

AsyncLetta

✓ from letta_client import AsyncLetta

Initialize the Letta client and list all agents.

import os from letta import Letta client = Letta(api_key=os.environ.get('LETTA_API_KEY', '')) agents = client.agents.list() print(agents)

Debug

Known issues

breakingThe client constructor changed from positional to keyword-only arguments in v1.x. Use `Letta(api_key=...)` instead of `Letta('...')`.

fix

Use keyword arguments: `Letta(api_key='...')`. For environment variable, omit api_key entirely.

affects: >=1.0.0

gotchaMethods that accept `model` or `model_id` may be case-sensitive. Use exact model identifiers from the Letta dashboard.

fix

Always use the exact model string as returned by `client.models.list()`.

affects: all

gotchaStreaming responses require `stream=True` parameter in message creation endpoints. Without it, the response is non-streaming and may hang if large.

fix

Add `stream=True` to create calls: `client.messages.create(..., stream=True)`.

affects: all

Upgrade

Version history

1.12.1latest on PyPI · released Jun 2, 2026

Audit

Dependencies

typing-extensionsrequiredUsed for extended type hints; minimum version required (>=4.7)

Agent activity

19 hits · last 30 days

node

Amazon

OpenAI (training)

Resources

githubhttps://github.com/letta-ai/letta-python ↗

packagehttps://pypi.org/project/letta-client/ ↗

homepagehttps://letta.com ↗