Key facts
| API compatibility | OpenAI-compatible /v1/chat/completions, /v1/embeddings and /v1/models |
| Migration effort | Change base_url and model name; keep your OpenAI SDK |
| Plane for this market | US plane, Virginia (us-east-1) is nearest for managed workloads; Brazilian residency needs a customer-managed deployment |
| Models | 30+ models behind one API, from free tiers to frontier reasoning |
| Pricing | Flat monthly self-serve plans, no per-token billing; see the live pricing page |
| Free tier | plugsky-micro and plugsky-lite with no credit card |
| Compliance alignment | The US plane is aligned with SOC 2 and HIPAA programs |
| Roadmap endpoints | Audio, images, moderation, files, batch, fine-tuning, assistants and responses are coming soon |
TL;DR
- Keep the OpenAI SDK; switch base_url to api.plugsky.com and map model names.
- Pin the US plane (Virginia) or deploy in your own network to satisfy the LGPD (Law No. 13,709/2018).
- Flat monthly pricing replaces per-token cost surprises on self-serve plans.
- Start free with plugsky-micro and plugsky-lite, or use the 14-day full-access trial for deeper testing.
- Where Plugsky does not replace OpenAI yet: audio, image and fine-tuning endpoints are still roadmap items.
How it works, step by step
- Inventory which OpenAI endpoints you call: chat, embeddings, streaming and tools.
- Create a Plugsky workspace and API key; the free plan needs no card.
- Point your OpenAI SDK at api.plugsky.com and map model names.
- Re-run your existing tests and evaluations against the same prompts.
- Pin the deployment or plane that satisfies the LGPD (Law No. 13,709/2018).
- Cut over gradually and keep the old base URL for rollback.
- Document the residency evidence Brazil reviewers will ask for.
Original data
Try it yourself
Open the OpenAI migration checker →
Why Brazil teams look for an OpenAI alternative
Brazilian teams weigh OpenAI alternatives against LGPD transfer mechanics and ANPD scrutiny, plus per-token costs that move with the exchange rate. Most evaluations start with cost predictability, followed by residency proof and model breadth; migration effort is usually the deciding factor, because rewrites are expensive.
Sovereign AI cloud is a set of choices, not a single product: where inference runs, where prompts and embeddings rest, who holds the encryption keys, and what evidence you can hand an auditor. Plugsky exposes those choices separately, so a regulated workload can be pinned while everything else stays on the default plane.
Migration: keep your OpenAI SDK and code
Because the API is OpenAI-compatible, migration is a configuration change rather than a rewrite: point the client at api.plugsky.com, map model names, and re-run your evaluation suite before cutting over.
- Inventory which OpenAI endpoints you actually call: chat, embeddings, streaming, tools.
- Create a Plugsky workspace and a scoped API key; the free plan needs no card.
- Point your SDK at api.plugsky.com and map model names.
- Re-run your existing tests and evaluations on the same prompts.
- Pin the deployment that satisfies the LGPD (Law No. 13,709/2018).
- Cut over gradually and keep the old base URL for rollback.
Data residency and compliance for Brazil
The LGPD (Law No. 13,709/2018) governs processing and international transfers, enforced by the ANPD. Financial institutions also follow Banco Central rules, including Resolucao CMN 4,658/2018 and Resolucao Conjunta no. 6/2023 on cloud use and processing location.
Brazilian teams that need processing in-country run a private endpoint, on-prem or air-gapped; others pin the US plane (Virginia) for proximity and accept a documented transfer basis. The US plane is aligned with SOC 2 and HIPAA programs.
Region-locked planes keep prompts, completions, embeddings and logs inside the chosen jurisdiction by architecture, and Enterprise agreements add BYOK, zero-knowledge mode and right-to-audit clauses.
What it costs, and what is still missing
Cost control is the headline: flat monthly self-serve pricing instead of per-token billing, a free plan with two models, and a 14-day full-access trial. Verify current plans on the live pricing page.
Portuguese-language evaluation sets matter: prompt quality on Portuguese support tickets rarely matches English benchmarks. Meet the mirror check: audio, image generation, moderation, files, batch, fine-tuning, assistants and responses are coming-soon endpoints rather than live today. Treat them as roadmap items when comparing vendors, and verify per-plane model availability before you commit.
Honest comparison
| Capability | Plugsky | OpenAI | Building in-house |
|---|---|---|---|
| API compatibility | OpenAI-compatible chat, embeddings and models endpoints | Native SDKs and API | Full rewrite |
| Pricing model | Flat monthly self-serve plans; see the live pricing page | Per-token billing | GPU plus operations cost |
| Model choice | 30+ models from one endpoint | OpenAI models only | You host each model |
| Data residency | Region-locked planes plus VPC, on-prem and air-gapped | US default with some enterprise options | Wherever you deploy |
| Migration effort | One-line base URL change | Not applicable | Months of engineering |
| Free tier | plugsky-micro and plugsky-lite, no card | Limited trial credits | Not applicable |
Frequently asked questions
Is Plugsky a drop-in replacement for OpenAI in Brazil?
Yes for the live surface: chat completions, streaming, JSON mode, function calling, embeddings, RAG and agent loops all work through OpenAI-shaped requests. You change the base URL and model name and keep your SDK.
How does data residency work for Brazil teams?
The LGPD (Law No. 13,709/2018) governs processing and international transfers, enforced by the ANPD. Financial institutions also follow Banco Central rules, including Resolucao CMN 4,658/2018 and Resolucao Conjunta no. 6/2023 on cloud use and processing location. Brazilian teams that need processing in-country run a private endpoint, on-prem or air-gapped; others pin the US plane (Virginia) for proximity and accept a documented transfer basis. Plugsky publishes planes in the EU (Frankfurt), GCC (UAE), APAC (Singapore) and US (Virginia), with Riyadh on Enterprise.
How is pricing different from OpenAI?
Self-serve Plugsky plans are flat monthly with no per-token metering, so forecasting is simpler; the free plan includes plugsky-micro and plugsky-lite, and a 14-day full-access trial exists. See the live pricing page for current plans.
Can we keep using the OpenAI SDK?
Yes. Plugsky exposes an OpenAI-compatible chat completions endpoint, so you change base_url to api.plugsky.com and map your model names while keeping existing SDK code.
How does the LGPD (Law No. 13,709/2018) affect migration?
The LGPD (Law No. 13,709/2018) governs processing and international transfers, enforced by the ANPD. Financial institutions also follow Banco Central rules, including Resolucao CMN 4,658/2018 and Resolucao Conjunta no. 6/2023 on cloud use and processing location. Brazilian teams that need processing in-country run a private endpoint, on-prem or air-gapped; others pin the US plane (Virginia) for proximity and accept a documented transfer basis.
Does streaming and function calling work?
Yes, streaming, JSON mode and function calling are live, as are embeddings, RAG and agent tool loops; check the docs for the per-model capability matrix.
What is not available yet?
Audio, image generation, moderation, files, batch, fine-tuning, assistants and responses endpoints are coming soon. Keep those workloads on their current provider until the status page marks them live.
How do we migrate and roll back safely?
Run both providers behind a routing flag: send a percentage of traffic to Plugsky, compare latency and quality on your own evaluation set, then switch fully. Rolling back is the same one-line base URL change in reverse.