Orchard Python Client
Use the Python client when your app can run Orchard in the same Python process. The client starts the local engine, loads models, and talks to the engine over IPC.
Chat Completion
from orchard.engine.inference_engine import InferenceEngine
MODEL = "google/gemma-4-E2B-it"
with InferenceEngine(load_models=[MODEL]) as engine:
client = engine.client()
response = client.chat(
MODEL,
[{"role": "user", "content": "Write one sentence about local AI."}],
temperature=0.0,
max_generated_tokens=64,
)
print(response.text)
Responses API
Use Responses when you want a typed output object, event streams, tools, structured output, or multimodal input.
from orchard.engine.inference_engine import InferenceEngine
MODEL = "google/gemma-4-E2B-it"
with InferenceEngine(load_models=[MODEL]) as engine:
client = engine.client()
response = client.responses(
MODEL,
input="Explain local inference in two sentences.",
temperature=0.0,
max_output_tokens=96,
)
print(response.output_text)
Async Client
import asyncio
from orchard.engine.inference_engine import InferenceEngine
MODEL = "google/gemma-4-E2B-it"
async def main() -> None:
async with InferenceEngine() as engine:
await engine.load_model(MODEL)
client = engine.client()
response = await client.achat(
MODEL,
[{"role": "user", "content": "Say hello from Orchard."}],
temperature=0.0,
max_generated_tokens=64,
)
print(response.text)
asyncio.run(main())
Common Parameters
| Parameter | Used by | Purpose |
|---|---|---|
temperature |
Chat and Responses | Sampling randomness |
top_p, top_k, min_p |
Chat and Responses | Token sampling controls |
max_generated_tokens |
Chat | Maximum generated tokens |
max_output_tokens |
Responses | Maximum generated tokens |
stream |
Chat and Responses | Return incremental deltas or events |
response_format |
Chat | Structured chat output |
text |
Responses | Structured Responses output |
tools, tool_choice |
Chat and Responses | Function calling |
reasoning_effort |
Chat | Native thinking effort for direct chat calls |
reasoning |
Responses | Native thinking effort for Responses calls |