Onboarding to a Cardano Data API: From Signup to First Call
What "onboarding" actually means for a blockchain data API
Every hosted data API puts the same four gates between you and your first byte of chain data: create an account, pick a plan, get a credential, make an authenticated request. The gates are universal. What differs between providers, and what actually decides how long onboarding takes, is what each gate demands of you and how much you can learn before you walk through it.
This matters most to the teams who evaluate tools hands-on before committing budget: dApp and DeFi builders wiring up token prices or wallet history for the first time, and engineering teams moving a product from testnet to mainnet who need a production-grade data layer under it. For them, time-to-first-call is not a vanity metric. It is the cost of finding out whether a provider fits at all.
TL;DR
Getting a Cardano API key takes minutes at any serious provider. What differs is what you learn on the way. Check four things: whether the docs are public before signup, what the plan gate costs you, how keys are scoped, and what the first response tells you about limits and errors. With Nexus the flow is verify your email, start the 7-day Builder trial, create a key in the dashboard, and send one request with an X-Api-Key header.
Gate | What to check | Nexus |
|---|---|---|
Docs | Public before signup, full spec, playground | Public, OpenAPI 3.1, live playground |
Plan | Free plan or trial, and what the trial exposes | No free plan. 7-day trial on Builder, $29/month |
Credentials | Auth mechanism, key scope, issuance speed |
|
First call | Rate-limit headers, distinguishable errors |
|
Where Nexus fits: one account and one base URL across Cardano, Bitcoin and Midnight, with an official TypeScript SDK and a Blockfrost-compatible surface for teams that are migrating.
The four gates, and what to check at each one
Gate 1: Can you read the docs before you sign up?
The cheapest evaluation is the one that needs no account. A provider that publishes its full endpoint reference publicly lets you answer the fit question, "does this API even return the data my product needs?", before you type an email address. Look for a complete OpenAPI spec rather than curated highlights, and ideally an interactive playground where you can see request and response shapes.
Most serious providers clear this bar. Blockfrost publishes its OpenAPI spec on GitHub and it has become the closest thing Cardano has to a de facto data standard. Koios documents its community-run API publicly. If a provider makes you sign up to see what the API returns, treat that as signal.
Gate 2: What does the plan gate cost you?
Providers split into two camps here. Some offer a permanent free tier: Blockfrost's Starter plan gives 50,000 requests per day at no cost, with no card, which is genuinely useful for hobby projects and long-running experiments. Others gate access behind a paid plan with a trial window, which trades away the permanent free home for full access to the paid surface during evaluation.
Neither model is better in the abstract. A free tier suits open-ended tinkering. A trial suits a scoped evaluation: you know what you need to test, you test it against the real paid product, and you decide. What you should check is whether the trial exposes the capabilities you are actually buying, or a reduced version of them.
Gate 3: How do credentials and auth work?
Three questions decide how much integration friction the credential model adds:
What is the auth mechanism? A single HTTP header is the least friction: no SDK required, every language's standard HTTP client works, and a curl one-liner proves the key is live.
What is the key scoped to? Some providers scope credentials per project, some per chain and network. Scoping affects how many credentials you manage as you grow, but a narrower scope also means a leaked mainnet key cannot touch anything else. Know the model before you design your config.
How fast is issuance? Key creation should be self-serve and instant from a dashboard, not a request to a sales team. If issuance involves a human, you are not onboarding to an API, you are onboarding to a vendor relationship. That is sometimes what enterprises want, but it should be a choice, not a surprise.
Gate 4: What does the first request teach you?
The first authenticated call is where the product starts telling you the truth about itself. Two things worth reading closely in that first response:
Rate-limit visibility. Does the API tell you where you stand? Headers like
X-RateLimit-Remainingon every response mean you can monitor quota programmatically from day one instead of discovering limits by hitting them.Error taxonomy. A provider that distinguishes "you are sending too fast" from "your monthly quota is exhausted" from "this capability is not on your plan", with different status codes and machine-readable bodies, is a provider you can build automated handling against. One that returns a generic 429 for everything leaves you guessing in production.
An evaluation that stops at "the key works" wastes the window. Point real traffic at the API: your query patterns, your peak request rates, your chained-transaction flows if you have them. Onboarding is fast everywhere the day nothing is wrong. What you are evaluating is what the API tells you when something is.
Where Nexus fits: the actual walk-through
Nexus is the full-stack data and transaction API for UTXO chains, built by Gero Labs: on-chain reads, market data across 11 Cardano DEXes, FIFO wallet P&L, server-side transaction building, and WebSocket streaming behind one REST surface, with Cardano, Bitcoin, and Midnight live today. Here is exactly what its four gates look like, stated plainly so you can hold us to it.
Before signup: the docs are public. The full endpoint reference, close to 200 endpoints on an OpenAPI 3.1 spec, is browsable with a live playground at nexus.gerowallet.io/docs. You can settle the fit question without an account.
Step 1: Create an account and verify your email at nexus.gerowallet.io.
Step 2: Choose a plan. Builder ($29/month) opens with a 7-day trial that exposes the full API surface. To be direct about the mechanics, because trial fine print is where onboarding friction likes to hide: there is no permanent free tier, adding a payment method is what starts the trial and unlocks key creation, and if you cancel before day 7 you pay nothing. You are evaluating the real paid product, not a demo of it.
Step 3: Create an API key from the dashboard. Issuance is self-serve and immediate. Each key is scoped to one chain and one network (a Cardano mainnet key is not a preprod key), and Builder includes 4 keys, so you can separate environments from the start. Keys start with nxs_ and are shown once, so store yours when you create it. Keys survive the trial-to-paid transition; nothing gets re-issued or breaks on day 8.
Step 4: Make the first call. The whole protocol is HTTPS plus an X-Api-Key header. No SDK is required and there is nothing else to install:
curl -H "X-Api-Key: nxs_your_api_key_here" \
https://nexus.gerowallet.io/api/blocks/latest
One mistake is worth avoiding on the first try. The key goes in the X-Api-Key header. Putting it in Authorization: Bearer returns 401, because that header is reserved for dashboard session tokens.
That is the full flow: verify, subscribe, create a key, curl. The only waiting is the email verification.
On the gate-4 questions, Nexus is built to be measured. Every metered response carries X-RateLimit-Limit, X-RateLimit-Remaining, and X-RateLimit-Reset headers. The two 429 cases are distinguishable (the monthly-quota one carries the quota headers and a reset_at body; the per-second one sends Retry-After: 1), and plan gating uses HTTP 402 with a machine-readable addon field. Builder meters at 2 million requests per month and 25 requests per second with a 250-token burst, and there is no overage billing: at the cap the API returns 429 until you upgrade, so the bill you modeled is the bill you get.
SDKs and migration paths
You do not need an SDK, but there is one. Nexus ships an official TypeScript client, @adlabs/nexus, installed with npm install @adlabs/nexus. Every other language works over plain REST, and the OpenAPI spec generates typed clients if you want them. If your consumer is an AI agent, the MCP server is the third route.
Arriving from Blockfrost. The same Nexus key works as a project_id against the Blockfrost-compatible surface, so existing Blockfrost SDK code is the shortest path in.
Arriving from Maestro. Maestro announced on 18 August 2026 that its developer API retires on 18 September 2026. The Cardano migration reference maps its endpoints to Nexus routes, and the Maestro alternatives guide covers the decision.
Two honest caveats so the picture is complete. Native providers for MeshJS and Lucid Evolution are not released yet. The Lucid Evolution provider is an open pull request upstream, so today you integrate through the SDK, the REST spec or the Blockfrost-compatible surface. And if your workload is pure on-chain reads with no market data, P&L, transaction building, or streaming, a cheaper single-purpose provider may fit better; the walk-through above only pays off if you need the stack it unlocks.
Related reading
Try the four gates yourself
The fastest way to test everything this article claims is to run the gates against your own workload. Browse the full endpoint reference and live playground at nexus.gerowallet.io/docs, then start the 7-day Builder trial at nexus.gerowallet.io and point a week of real traffic at it. For Dedicated infrastructure, custom SLAs, or compliance documentation, reach the team through the support page.