NemoRouter is an OpenAI-compatible LLM gateway. Migrating from OpenRouter, OpenAI, or Portkey means changing the base URL and the API key — nothing else. Your SDKs and frameworks work unchanged, and you land on 20+ models, one bill, and built-in guardrails.
The entire change
base_url + api_key — your logic is untouched
No SDK swap, no code rewrite
Annual plan — vs OpenRouter 5-15% per call
Google Vertex AI today, more shipping
Whichever gateway or provider you start from, the diff is the same shape: repoint the base URL, swap the key. Your prompts, parameters, streaming, and error handling are unchanged.
# Before — OpenRouter
client = OpenAI(
base_url="https://openrouter.ai/api/v1",
api_key="sk-or-...",
)
# After — NemoRouter (only two lines change)
client = OpenAI(
base_url="https://api.nemorouter.ai/v1",
api_key=os.environ["NEMOROUTER_API_KEY"],
)# Before — OpenAI direct
client = OpenAI(
base_url="https://api.openai.com/v1",
api_key=os.environ["OPENAI_API_KEY"],
)
# After — NemoRouter (only two lines change)
client = OpenAI(
base_url="https://api.nemorouter.ai/v1",
api_key=os.environ["NEMOROUTER_API_KEY"],
)# Before — Portkey gateway
client = OpenAI(
base_url="https://api.portkey.ai/v1",
api_key=os.environ["PORTKEY_API_KEY"],
)
# After — NemoRouter (only two lines change)
client = OpenAI(
base_url="https://api.nemorouter.ai/v1",
api_key=os.environ["NEMOROUTER_API_KEY"],
)The same two-line pattern applies to the Node, Go, Java, Ruby, and PHP SDKs — the canonical, runnable version of the “after” snippet is below.
This is the canonical NemoRouter client — generated from the same SDK examples the playground and dashboard use. Set NEMOROUTER_API_KEY in your environment and this runs as-is.
pip install openai| 1 | # Cache: enabled (org default). Pass nemo_cache: false to skip. |
| 2 | from openai import OpenAI |
| 3 | import os |
| 4 | |
| 5 | client = OpenAI( |
| 6 | api_key=os.environ["NEMOROUTER_API_KEY"], |
| 7 | base_url="https://api.nemorouter.ai/v1", |
| 8 | ) |
| 9 | |
| 10 | response = client.chat.completions.create( |
| 11 | model="gemini-2.5-flash", |
| 12 | temperature=1, |
| 13 | max_tokens=1024, |
| 14 | top_p=1, |
| 15 | messages=[ |
| 16 | {"role": "user", "content": "Hello! What models do you support?"}, |
| 17 | ], |
| 18 | extra_body={ |
| 19 | # "nemo_cache": False, # Uncomment to skip cache |
| 20 | }, |
| 21 | ) |
| 22 | |
| 23 | print(response.choices[0].message.content) |
Base URL is always https://api.nemorouter.ai/v1.
Run NemoRouter alongside your current setup if you want — repoint a slice of traffic, verify, then move the rest. There is no migration fee and no lock-in.
Sign up at nemorouter.ai, add credits, and generate a NemoRouter virtual key (sk-nemo-…).
Set your client base_url to https://api.nemorouter.ai/v1 — the same OpenAI-compatible path your code already calls.
Replace your old provider/gateway key with the NemoRouter virtual key. That is the second and final line.
Send a test request, confirm the response shape, then shift the rest of your traffic when you are ready.
The mechanics are identical; the context is not. Here is what specifically changes — and what you gain — depending on what you run today.
OpenRouter is OpenAI-compatible, so the migration is purely a base-URL and key swap. The upside is on the bill and the controls.
Calling OpenAI directly means one provider and your own key management. Routing through NemoRouter keeps the OpenAI API but unlocks the catalog.
Portkey is also a gateway, so you are already used to repointing a base URL. NemoRouter keeps that model and folds the add-ons into every tier.
Comparing in detail? See Nemo vs OpenRouter and Nemo vs Portkey.
The migration is trivial on purpose. The point is everything that happens once your traffic flows through a managed gateway.
0% platform fee on Tier 3 annual, 2% on Tier 2 monthly, 4% pay-as-you-go. The fee is charged on top of credits — you always keep 100% of what you buy.
PII redaction, prompt-injection detection, and secret scanning run on every request — on every tier, at no extra cost.
Owner / admin / member / viewer roles, per-team budgets, and per-key spend — RLS-enforced tenant isolation under the hood.
Every feature is unlocked on every plan. Tiers only change the platform fee and RPM/TPM guarantees — never what you can use.
Common questions teams ask before switching gateways.
Two lines · zero migration fee
Create a free account, generate a key, repoint your base URL. Keep your code, your SDKs, and your frameworks — gain one bill, guardrails, and lower fees.
Moving a large workload? Email sales@nemorouter.ai for assisted, parallel-run migration support.