Protecto alternative: Nonimo Gateway on your own server
· Written and maintained by Nonimo
Protecto is an API-first data protection platform that masks personal data before text reaches a large language model and restores it afterwards through an explicit Unmask API. Its self-serve plans are priced by API calls, from $250 a month for 5,000 calls, and running it on your own servers is an Enterprise option.
Nonimo Gateway runs on your own servers. See plans and get in touch.
People looking for a Protecto alternative usually want one of two things: to run it on their own servers without an Enterprise contract, or to change one URL instead of wrapping every request in mask and unmask calls. This guide sets out what Protecto publishes and where Nonimo Gateway fits.
Protecto pricing: the published ladder
Protecto’s pricing page lists five plans, using the $ symbol without naming the currency.
| Plan | Price | Calls included |
|---|---|---|
| Free Trial | $0 for 14 days | 1,000 API calls |
| Starter | $250 a month | 5,000 calls a month |
| Growth | $500 a month | 12,000 calls a month |
| Scale | $750 a month | 16,000 calls a month |
| Enterprise | Custom | Contact sales |
It publishes its prices and offers a trial without a sales call. It is built around permission-controlled unmasking, where you decide who may see the real values.
Two things to check before you sign. Masking and unmasking are separate operations in Protecto’s documentation, so ask how a round trip counts against the quota. And rate limits apply per namespace and per subscription tier, so price your real volume, not the pilot’s.
Where Protecto may not fit
- Self-hosting is Enterprise. The pricing page lists dedicated VPC, on-premises and self-hosted deployment under Enterprise; Starter to Scale are presented as the SaaS API. If text must stay on your servers, the published prices do not apply. Hardware requirements are not published.
- It is an API, not a proxy you host. Protecto documents the GenAI case as mask, call the LLM, then optionally restore. We found no documentation for an OpenAI-compatible proxy as of 20 October 2026, although the homepage promotes a “Privacy Gateway for AI Chat”.
- Streaming is not documented. Neither are tool calls. Your code handles any streamed response itself.
- Seven languages. The supported languages page lists English, German, French, Spanish, Dutch, Arabic and Japanese. Italian and Portuguese are not on it.
Reversible replacement is pseudonymisation, and the data is still personal data in law; our guide on deidentified versus anonymised explains why.
Protecto and Nonimo Gateway side by side
| Protecto (self-serve plans) | Nonimo Gateway | |
|---|---|---|
| What you pay for | A monthly quota of API calls | Plans on the Nonimo Gateway page |
| Where it runs | Protecto’s SaaS; VPC and on-premises under Enterprise | On your own servers |
| How your code uses it | Mask call, model call, unmask call | Change base_url in your OpenAI client |
| Restoring real values | Explicit Unmask API, permission-controlled | Automatic in the response |
| Streaming | Not documented | Real values back in the answer |
| Coverage | Seven languages listed | National documents and numbers for each country it serves |
A per-call SaaS plan suits low or uneven volume. If the text has to stay on your own machines, the comparison is with Protecto’s Enterprise terms, not the published ladder. Our overview of PII redaction software compares the other pricing shapes in this market.
How Nonimo Gateway does it
Nonimo Gateway is a private proxy on your own servers. It exposes an OpenAI-compatible API, so the change in your code is the address:
import os
from openai import OpenAI
client = OpenAI(
base_url="https://gateway.your-company.internal/v1", # your Nonimo Gateway
api_key=os.environ["OPENAI_API_KEY"], # your own OpenAI key
)It replaces personal data in what goes out and restores the real values in the response. It recognises the national documents and numbers of each country it serves, among them the Spanish DNI, NIE and NIF, the German Steuer-ID, the French NIR, the Italian codice fiscale, the Brazilian CPF and the Mexican RFC and CURP. If you route calls through LiteLLM, there is a plugin; see LiteLLM PII guardrails.
Whichever way you go, write down what may and may not be sent first; our AI policy template gives you the headings. For other comparisons, see the Limina (formerly Private AI) alternative and the Skyflow alternative.
If you want the text to stay on your own server and the real values back in the answer, Nonimo Gateway is built for that. See plans and get in touch on the Gateway page.
Sources
Checked 20 October 2026.
- Protecto, homepage. Positioning, deployment in SaaS, VPC or on-premises, the privacy gateway promotion.
- Protecto, pricing. The five plans, prices and call quotas, and what Enterprise includes.
- Protecto Docs, index. API-first framing, identify and mask, no proxy documentation.
- Protecto Docs, unmasking. Unmasking is explicit, permission-controlled and never automatic.
- Protecto Docs, GenAI and LLM prompts. Mask before the LLM, optionally restore after.
- Protecto Docs, rate limits. Limits per namespace and tier.
- Protecto Docs, supported languages. The seven languages.
This page is for information. Product details of other vendors were checked on their official pages on 20 October 2026 and can change.
Nonimo is the software that does this on your own computer: it redacts client names and IDs before your text reaches ChatGPT. No account needed, and the app does it without your files leaving your machine.
Common questions
How much does Protecto cost?
Protecto publishes a ladder on its pricing page: Starter at $250 a month for 5,000 calls, Growth at $500 a month for 12,000 calls, and Scale at $750 a month for 16,000 calls. Enterprise is custom. A free trial gives 14 days and 1,000 API calls.
Can Protecto restore the original values after the LLM answers?
Yes. Protecto documents an Unmask API that restores original values from tokens. Its documentation describes unmasking as always explicit, permission-controlled and never automatic, so your code makes the call when it needs the real values.
Can I self-host Protecto?
Protecto states it can run in SaaS, a private VPC or fully on-premises. Its pricing page lists dedicated VPC, on-premises and self-hosting under the Enterprise plan; the self-serve paid tiers are presented as the SaaS API. Hardware requirements are not published.
Which languages does Protecto support?
Its documentation lists seven: English, German, French, Spanish, Dutch, Arabic and Japanese. Italian and Portuguese are not on that list as of 20 October 2026.
Does Protecto work as an OpenAI-compatible proxy?
Not according to its documentation as of 20 October 2026. The homepage promotes a privacy gateway for AI chat, but the technical documentation describes an API you call to mask and unmask. Streaming and tool calls are not documented.
How is Nonimo Gateway sold compared with Protecto?
Protecto's self-serve plans are SaaS plans with a monthly quota of API calls, and running it on your own servers is an Enterprise option. Nonimo Gateway runs on your own servers and uses your own OpenAI key. Its plans and prices are on its page.