Python SDK

A thin, dependency-free client for the /analyze endpoint. One call tests an app and returns root-caused bugs — each failing case resolved to its backend cause via trace_id.

bash

pip install upliftr        # stdlib only, no runtime deps

Source: sdk/python in the repo. Point it at any Upliftr deployment — your own self-hosted engine, or an OEM's.

Usage

python

from upliftr import Upliftr

up = Upliftr("https://YOUR-UPLIFTR", token="tw_...")

result = up.analyze("https://staging.myapp.com")

if result.passed:
    print("all good")
else:
    for bug in result.bugs():
        print(f"❌ {bug.title} ({bug.status})")
        print(f"   why:            {bug.why}")
        print(f"   backend cause:  {bug.backend_cause}   [{bug.confidence}]")
        print(f"   suggested fix:  {bug.suggested_fix}")
        print(f"   trace_id:       {bug.trace_id}")

Reference

`Upliftr(base_url, token, timeout=900)`

Point at any Upliftr deployment (no /api/v1 suffix). token is a tw_… org-scoped API token. timeout is in seconds (a sync analyze can run minutes).

`up.analyze(app_url, *, ...) -> AnalyzeResult`

Arg	Meaning
`app_url`	the app to test
`auth`	`{username, password, login_path}` to log in first
`exhaustive`	deeper coverage (default `True`)
`file_issues`	`"github" \| "gitlab" \| "jira" \| "native"`
`log_backend`	per-request server-log config (see server logs)
`anthropic_key`	per-request BYOK key
`mode`	`"sync"` (wait) or `"async"` (returns a `run_id`)
`timeout_s`	sync wait budget

`AnalyzeResult`

.run_id, .status, .summary, .passed, .raw (full JSON), and:

.bugs() → a list of Bug, one per failing/broken case, each with: .title, .status, .why, .category, .backend_cause, .suggested_fix, .confidence, .trace_id, .evidence.

Errors

Any non-2xx or transport failure raises UpliftrError with the status + detail.

python

from upliftr import Upliftr, UpliftrError
try:
    up.analyze("https://x")
except UpliftrError as e:
    print("analyze failed:", e)

Python SDK ​

Usage ​

Reference ​

Upliftr(base_url, token, timeout=900) ​

up.analyze(app_url, *, ...) -> AnalyzeResult ​

AnalyzeResult ​