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 97+ models, one bill, and built-in guardrails.
No migration fee · run both in parallel during cutover
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 — Nemo Router (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 — Nemo Router (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 — Nemo Router (only two lines change)
client = OpenAI(
base_url="https://api.nemorouter.ai/v1",
api_key=os.environ["NEMOROUTER_API_KEY"],
)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-lite", |
| 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) |
The canonical post-migration client, generated from the same SDK examples the playground and dashboard use. Base URL is always https://api.nemorouter.ai/v1.
Run Nemo Router 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 Nemo Router 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 Nemo Router 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.
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 Nemo Router keeps the OpenAI API but unlocks the catalog.
Portkey is also a gateway, so you are already used to repointing a base URL. Nemo Router 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 — lower fees, guardrails, real team management, and zero feature gating.
Day 0 — before
Per-call gateway fees, no controls
Spend untracked per team · guardrails are your problem
Day 0 — after two lines
Same code, managed gateway
Guardrails, budgets, and audit logs on every request
Every day after
One bill, exact costs
Credits reserve + settle from the provider-reported cost
Most teams migrate in under five minutes. Nemo Router exposes an OpenAI-compatible REST API, so you change two things — the base URL and the API key — and your existing code, SDKs, and frameworks keep working without modification.
No. Nemo Router speaks the OpenAI API. If you use the OpenAI Python or Node SDK, LangChain, LlamaIndex, the Vercel AI SDK, or any OpenAI-compatible client, you only repoint the base URL. Request and response shapes are identical.
Yes. OpenRouter is also OpenAI-compatible, so the migration is a base-URL and key swap. Nemo Router additionally gives you built-in guardrails, team management, audit logs, and a lower platform fee — 0% platform fee on Tier 3 annual ($1,200+/yr) versus OpenRouter's published per-call fee.
Yes. Replace api.openai.com with the Nemo Router endpoint and use a Nemo Router virtual key. You keep one OpenAI-style API but gain access to multiple model providers, built-in guardrails, per-team budgets, and real-time cost tracking — without managing provider keys yourself.
You select models by their Nemo Router model id (for example gemini-2.5-flash). The catalog and exact ids are listed on the models page; switching models is a one-string change with no SDK swap.
No migration fee. You can run Nemo Router alongside your current gateway during cutover — point a percentage of traffic at the new base URL, verify, then move the rest. Nemo Router only charges the platform fee on credits you actually buy.
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.