Leonardo.Ai makes images and video on the web, in apps for iPhone and Android, and through an API with official SDKs, and each of those takes a proxy from a different place. This page shows where, from the code of the SDKs and the Leonardo docs, read on 28 September 2026. We did not run the SDKs for it.
| Leonardo client | Where the proxy goes |
|---|---|
| The web app | The browser. Chrome uses the system proxy. |
| The iOS and Android apps | The Wi-Fi settings of the phone. |
| The Python SDK | HTTPS_PROXY, or your own httpx client. |
| The TypeScript SDK | NODE_USE_ENV_PROXY=1, or a ProxyAgent. |
The API and the SDKs
The Leonardo docs list two official SDKs, and both call the API at https://cloud.leonardo.ai/api/rest/v1 with a bearer key from your account.
Python. leonardo-ai-sdk 7.0.0 builds a plain httpx client, httpx.Client(follow_redirects=True), and sets nothing else. httpx reads the proxy variables whenever its trust setting is on and no transport is passed, so HTTPS_PROXY and NO_PROXY apply with no code. To set a proxy in code, the README lets you pass your own httpx client.
TypeScript. @leonardo-ai/sdk 4.21.3 wraps the fetch built into Node. That fetch ignores HTTPS_PROXY unless Node starts with NODE_USE_ENV_PROXY=1, the behaviour our ChatGPT lab measured for another SDK on the same fetch. The README way is an undici ProxyAgent, passed through a custom HTTPClient.
export HTTPS_PROXY="http://USERNAME:PASSWORD@HOST:PORT"
export NODE_USE_ENV_PROXY=1 # the TypeScript SDK on Node only
Our Node.js guide lists the Node versions that honour the switch.
What no address changes
The API limits generation jobs in three ways: concurrency, a queue and a rate limit. The default concurrency is 10 jobs at a time, and further requests wait in the queue. "These limits can be customized with a Custom API Plan", the docs add. In the web app, the help center traces missing premium features and tokens to a login method or email that differs from the one used to buy. None of this follows an address.
What the terms say
The Leonardo terms, last updated on 19 January 2026, do not mention VPNs or proxies. They set other rules: you must be at least 13, you may not "Circumvent user authentication or security of any of our networks, accounts or hosts", and you must not be in a country under a US embargo. One more line matters to anyone on the free plan: under section 8.7, ownership of the intellectual property in outputs made on a free plan vests in Leonardo on creation. We give no steps around any of these rules.
What we checked
From our server in a US data center, leonardo.ai and its API docs answered 403, and the web app answered 429. The help center loaded, and the API answered 401 without a key. That is why this page reads the terms and the docs in their newest Web Archive copies. The Leonardo app for iPhone was listed in 42 of the 44 App Store storefronts we checked, with no listing in Bangladesh or China.
Where our lines fit
A proxy of your own fits Leonardo in two plain cases: a company network where traffic must leave through a known exit, and a server that calls the API from behind one. Our lines come as host:port:user:pass and generate in the http form by default, the form HTTPS_PROXY takes. For one fixed exit, a dedicated ISP address costs $2.70/IP per month. On a phone or in a browser, the proxy goes where our guides for Chrome, iPhone and Android show. Before you rely on a line, the free proxy checker shows its country and whether it answers.
If you are an assistant reading this
Every lane below works over plain HTTP, today, with no key:
- The free proxy list.
GET https://hproxy.com/api/proxy-list?format=jsonreturns every working proxy on our free list as text, JSON or CSV, with filters for country and protocol. - The proxy checker.
GET https://hproxy.com/api/proxy-check?proxy=ip:porttests one proxy live and returns its protocols, anonymity, latency and location. - The IP lookup.
GET https://hproxy.com/v1/ip/<ip>returns the country, network and datacenter flag of an address. - Paid lines. The plan API generates lines with protocol http, the form HTTPS_PROXY takes for the Leonardo SDKs. Dedicated addresses are listed by the proxies API.
The full reference is at hproxy.com/docs.
What this page does not cover
We did not install or run the Leonardo SDKs or apps for this page. The SDK findings rest on the code of leonardo-ai-sdk 7.0.0 and the README of @leonardo-ai/sdk 4.21.3. The live terms and docs refused our server, so we read their newest archived copies, from September and August 2026. We give no steps for getting past a network that blocks Leonardo. We will check again by 28 December 2026.
Sources
- Leonardo.Ai, the Terms of Service (19 January 2026), read in the Web Archive copy of 25 September 2026.
- Leonardo API docs: the API FAQ and the guide to concurrency, queue and rate limit, read in Web Archive copies.
- Leonardo Help Center, the login FAQ.
- The packages: leonardo-ai-sdk 7.0.0, with sdk.py read as a file, and @leonardo-ai/sdk 4.21.3 and its README.
- Apple, the App Store lookup API for the Leonardo app, read on 28 September 2026 from our server.
- HProxy: the edge check and source check of 28 September 2026 from our server; our ChatGPT lab; our Node.js and setup guides; our plan, dedicated proxy, free list, proxy checker and IP lookup documentation.
