vexp-sdk (Python)

Thin client for the vexp SDK — a self-hosted, machine-to-machine code-intelligence engine. It gives your agents, PR bots, and pipelines type-accurate context and cross-repo blast-radius over hundreds of repositories, on your own infrastructure. Your code never leaves your network.

It is the same engine that powers vexp for AI coding agents — where, measured on full paired agent sessions, it resolves more tasks while cutting context replay by double digits — packaged for fleet-scale, machine-to-machine use.

This package is only the client. It talks to a vexp gateway you run yourself (see the docs); it needs the gateway's URL and a bearer token. It is dependency-free (standard library only), so it installs and runs anywhere, including air-gapped CI.

Install

pip install vexp-sdk

Quickstart

from vexp_sdk import VexpClient

vx = VexpClient("https://vexp.your-company.internal", token="...")

# Context + impact for a task
res = vx.run_pipeline("why does checkout fail", repos=["payments-gateway"])
for p in res["pivots"]:
    print(p["repo"], p["file_path"], p["score"])

# Cross-repo blast-radius of a symbol
impact = vx.impact_graph("payments.Checkout.create_session")

# Drive a re-index after a deploy
vx.index_incremental(repo="payments-gateway", mode="incremental")

API

VexpClient(base_url: str, token: str | None = None, timeout: float = 60.0)

health() -> dict                                       # gateway liveness + shard count
run_pipeline(task: str, **kwargs) -> dict              # context + impact + memory
context_capsule(query: str, **kwargs) -> dict          # retrieval only
impact_graph(symbol_fqn: str, **kwargs) -> dict        # cross-repo blast-radius
index_incremental(repo=None, files=None, mode="incremental") -> dict
index_status() -> dict
call(tool: str, **params)                              # generic tool call

run_pipeline returns a dict with pivots (ranked code locations) and an optional compressed output. Non-2xx responses raise VexpError.

Requirements

• A running vexp gateway (self-hosted) reachable at base_url.
• A bearer token for the gateway, and an SDK license.jwt mounted in the gateway (issued with your plan).

Full deployment and API reference: https://vexp.dev/docs/sdk

License

Proprietary. The client is free to install and use against a licensed gateway; the engine itself is license-gated. See https://vexp.dev/sdk.

_{Keywords: code intelligence SDK · self-hosted code context · cross-repo impact analysis · blast radius API · monorepo dependency graph · AI agent context · PR review automation · on-premise code graph · Python client}