lifi-py 0.1.0

Python client for the LI.FI API: bridge + DEX aggregator for any-to-any cross-chain swaps.

lifi-py

A typed Python client for the LI.FI API — the bridge + DEX aggregator for any-to-any cross-chain swaps and bridges across 74 chains.

The killer feature: get_quote() returns a ready-to-sign transactionRequest directly — one call, no separate assemble step.

• Sync (LifiClient) and async (AsyncLifiClient) clients on top of httpx.
• All responses parsed into pydantic v2 models with full type hints (py.typed).
• Built-in rate-limit awareness: reads the ratelimit-* headers, throttles proactively, and retries 429s with backoff.
• Optional web3.py integration for signing/sending the returned transaction.

Install

pip install lifi-py
# with web3 signing helpers:
pip install 'lifi-py[exec]'

Quickstart — cross-chain quote with a ready-to-sign tx (keyless, <5 lines)

from lifi_py import LifiClient

quote = LifiClient().get_quote(
    from_chain=42161, to_chain=8453,                                  # Arbitrum -> Base
    from_token="0xaf88d065e77c8cC2239327C5EDb3A432268e5831",          # USDC (Arbitrum)
    to_token="0x833589fCD6eDb6E08f4c7C32D4f71b54bdA02913",            # USDC (Base)
    from_amount="1000000", from_address="0x47E2D28169738039755586743E2dfCF3bd643f86",
)
print(quote.estimate.to_amount, quote.transaction_request.to)  # ~995000, 0x1231DEB6...

quote.transaction_request is an EVM transaction ready to sign and broadcast.

Signing & sending (optional, [exec] extra)

tx = quote.transaction_request.as_web3_tx()   # hex fields decoded to ints, web3-shaped
signed = account.sign_transaction({**tx, "nonce": w3.eth.get_transaction_count(account.address)})
tx_hash = w3.eth.send_raw_transaction(signed.raw_transaction)

Tracking a bridge to completion

client = LifiClient()
status = client.poll_status(tx_hash="0x...", bridge=quote.tool)  # polls until DONE/FAILED
print(status.status, status.is_done)

get_status() does a single check; poll_status() polls with exponential backoff until the transfer reaches a terminal state (or raises TimeoutError).

Async

import asyncio
from lifi_py import AsyncLifiClient

async def main():
    async with AsyncLifiClient() as client:
        quote = await client.get_quote(...)

asyncio.run(main())

Endpoints

Method	API endpoint
get_quote(...)	GET /quote — best single-step route + transactionRequest
get_status(...) / poll_status(...)	GET /status — track a transfer
get_routes(...)	POST /advanced/routes — multiple candidate routes
get_step_transaction(step)	POST /advanced/stepTransaction — calldata for one step
get_chains()	GET /chains
get_tokens()	GET /tokens
get_tools()	GET /tools — bridges + exchanges
get_connections(...)	GET /connections

Authentication & rate limits

An API key is optional — every endpoint above works keyless. A key only raises your rate limits:

client = LifiClient(api_key="your-key")                 # sent as `x-lifi-api-key`
client = LifiClient(api_key="k", api_key_header="x-custom-key")  # configurable header
client = LifiClient(base_url="https://staging.li.quest/v1")      # configurable base URL

The client reads the API's ratelimit-limit / ratelimit-remaining / ratelimit-reset response headers (exposed via client.rate_limit). When the quota is exhausted it pauses before the next request; on a 429 it retries with backoff respecting the reset window, raising LifiRateLimitError once retries are exhausted.

Configuration

Argument	Default	Description
api_key	None	Optional; raises rate limits.
base_url	https://li.quest/v1	API base URL.
api_key_header	x-lifi-api-key	Header name for the key.
timeout	30.0	Per-request timeout (seconds).
max_retries	3	Max 429 retries before raising.
http_client	None	Inject your own httpx.Client / AsyncClient.

Development

pip install -e '.[dev,exec]'
ruff check .
mypy
pytest                 # unit tests (mocked HTTP, no network)
pytest -m integration  # live smoke tests against the keyless API

License

MIT