langchain-python

LangChain (Python) on Inferit

ChatOpenAI from langchain-openai, pointed at Inferit's OpenAI-compatible base URL with your API key. Guide: https://inferit.cc/docs/guides/langchain

Testnet demo: test tokens, no monetary value. Requests are paid in ITC (Inferit Test Credit).

File	What it does
chat.py	One request with an API key (max_tokens 16)
requirements.txt	langchain-openai
smoke.sh	Makes a .venv here, installs the requirements and runs chat.py

Run it

You need Python 3.10 or newer and an API key (ik_...) from https://inferit.app/buy.

export INFERIT_API_KEY='ik_...'
bash smoke.sh                       # or PYTHON=python3.12 bash smoke.sh

By hand:

python3 -m venv .venv && .venv/bin/pip install -r requirements.txt
INFERIT_MODEL=qwen/qwen-2.5-14b-instruct .venv/bin/python chat.py

Environment

Variable	Default
INFERIT_API_KEY	required
INFERIT_BASE_URL	https://api-production-c74b9.up.railway.app/v1 (the hosted testnet API)
INFERIT_MODEL	auto in chat.py; smoke.sh picks one with ../pick-model.sh
PYTHON	python3 (smoke.sh only)

reply.response_metadata["model_name"] names the model that answered, which matters with auto.

verified 2026-10-13: langchain-openai 1.7.1 on Python 3.12, model qwen/qwen-2.5-14b-instruct, max_tokens 16: "Hello, how can I assist?" (model: qwen/qwen-2.5-14b-instruct).