timestep 2026.0.9

Multi-model provider implementations for OpenAI Agents, supporting both OpenAI and Ollama models

Timestep (Python)

Python bindings for the Timestep Agents SDK. See the root README.md for the full story; this file highlights Python-specific setup.

Install

pip install timestep

Prerequisites (Python)

• OPENAI_API_KEY
• Default storage uses PGLite via Node: install Node.js and @electric-sql/pglite (npm install -g @electric-sql/pglite or local install).
• For better performance use Postgres: set TIMESTEP_DB_URL=postgresql://user:pass@host/db. If you must stay on PGLite, keep a long-lived Node/Deno sidecar that holds a PGlite connection instead of spawning Node per query.

Quick start

from timestep import run_agent, RunStateStore, consume_result
from agents import Agent, Session

agent = Agent(model="gpt-4o")
session = Session()
state_store = RunStateStore(agent=agent, session_id=await session._get_session_id())

result = await run_agent(agent, input_items, session, stream=False)
result = await consume_result(result)

if result.interruptions:
    await state_store.save(result.to_state())

Cross-language resume

Save in Python, load in TypeScript with the same session_id/run_id and RunStateStore.load().

Model routing

Use MultiModelProvider if you need OpenAI + Ollama routing:

from timestep import MultiModelProvider, MultiModelProviderMap, OllamaModelProvider

provider_map = MultiModelProviderMap()
provider_map.add_provider("ollama", OllamaModelProvider())
model_provider = MultiModelProvider(provider_map=provider_map)

Documentation

Full docs: https://timestep-ai.github.io/timestep/