Archive DataEthereumEVMHistorical DataAPI

Alternatives to Running an Archive Node for Historical EVM Data

An Ethereum archive node needs terabytes of fast disk and weeks of upkeep. Hosted archive RPC, indexers, decoded data APIs and Parquet datasets get you the same history. What each option costs and when to use it.

By Divyasshree · Blockchain Intelligence

Cover Image for Alternatives to Running an Archive Node for Historical EVM Data

The short answer

You rarely need to run your own archive node. Pick the option that matches the question you ask most:

You need	Best fit	Examples
Old state through standard JSON-RPC calls	Hosted archive RPC	QuickNode, Alchemy, Chainstack
One protocol's events, shaped your way	An indexer or subgraph	The Graph, Goldsky
Decoded trades, transfers, balances and prices across chains	A decoded data API	Bitquery
Years of history to load into your own warehouse	Bulk files or cloud delivery	Bitquery Data Store, public cloud datasets

The rest of this guide explains each option and what it costs.

Why teams avoid running an archive node

An archive node keeps every past state of the chain, so you can ask what any account held at any block. That is exactly what backtests, tax reports and investigations need.

Running one is the hard part. An Ethereum archive node needs several terabytes of fast SSD and a sync that takes days to weeks. It then needs monitoring, client upgrades and a second copy for failover. And it only covers one chain. A team on five EVM chains runs five of them.

Even a healthy archive node only returns raw data. To get "every swap this wallet made in 2025" you still pull logs block by block, decode each event and add a USD price. That pipeline is usually the real cost.

Option 1: hosted archive RPC

Providers such as QuickNode, Alchemy and Chainstack run archive nodes for you and sell access to the same JSON-RPC methods.

Good for: existing code that already calls eth_call or eth_getBalance at an old block.
Watch for: archive access is often a paid feature, and wide eth_getLogs ranges are capped. On 29 October 2026, Chainstack's free plan had no archive data, while QuickNode listed archive data on every plan.
You still build: the decoding, pricing and indexing on top.

Monthly prices are compared in blockchain data API pricing.

Option 2: an indexer or subgraph

Indexers such as The Graph and Goldsky read the chain and store only the events you define, in the shape you want.

Good for: one protocol or a small set of contracts you care about deeply.
Watch for: you write and maintain the mapping code, and a new question often means a re-index from the start block.

See Bitquery vs The Graph for how this compares with a pre-indexed API.

Option 3: a decoded data API

A decoded data API has already indexed the chain. You ask for trades, transfers, balances, holders, events or calls, and the answer comes back decoded with USD values where they apply.

Bitquery covers this through GraphQL, WebSocket streams and Kafka. On self-service plans, recent data comes in a rolling window. Complete history on a core chain is a flat-rate archive add-on:

Archive add-on	Monthly	Includes
Ethereum, BSC, Base, Arbitrum, Polygon or Optimism, trading data	$150 per chain	Complete trades, prices and OHLC
Ethereum, BSC, Base, Arbitrum, Polygon or Optimism, transfers	$150 per chain	Complete transfers, balances and holders
All 8 EVM chains, trading data	$600	The same, as a bundle
All 8 EVM chains, transfers	$600	The same, as a bundle

On EVM chains, both packs also carry events and calls. Traces are outside the packs and ship as an S3 export instead. Prices billed yearly are lower. Every add-on is listed on the pricing page.

Good for: cross-chain questions, wallet and token history, and anything that needs prices next to the trades.
Watch for: you query Bitquery's schema rather than raw RPC. There is no eth_call replay of contract state.

Option 4: bulk files and cloud delivery

When you need years of history inside your own warehouse, download it once instead of calling an API.

Parquet files: the Bitquery Data Store sells chain datasets such as Ethereum as partitioned Parquet files you can load into DuckDB, Spark or ClickHouse. Every dataset has a free sample file.
Delivered to your cloud: Datashares and the blockchain data lake keep tables updated in Snowflake, BigQuery or S3.
Public cloud datasets: Google BigQuery and AWS host free public crypto datasets. They cover a handful of chains with raw tables, and you pay for the queries you run.

How to choose

List your top three questions. "What did this wallet hold on 1 January?" and "every Uniswap swap in 2024" need different tools.
Count your chains. One chain and one protocol suits an indexer. Several chains suit a decoded API or bulk files.
Decide where the data lives. Querying on demand suits an API. Joining with your own tables suits Parquet or cloud delivery.
Price the full pipeline. Add the engineering time to decode and price raw data to the monthly bill.

FAQ

What are the best alternatives to running an archive node?

Hosted archive RPC from providers such as QuickNode, Alchemy or Chainstack; indexers such as The Graph or Goldsky; decoded data APIs such as Bitquery; and bulk Parquet datasets or cloud delivery for warehouse use.

Can I get historical token balances without an archive node?

Yes. A decoded data API returns balances and balance changes directly. On Bitquery, complete balance and holder history on an EVM chain comes with the $150 a month transfers add-on.

Where can I download full Ethereum history as Parquet files?

The Bitquery Data Store sells Ethereum datasets as partitioned Parquet files, with a free sample for each. Google BigQuery and AWS also host public Ethereum tables that you query in place.

How much does historical blockchain data cost?

Hosted archive RPC ranges from free trials to plans around $49 a month and up. On Bitquery, complete history for Ethereum or another major EVM chain is $150 a month per dataset on top of a plan, or $600 a month for all 8 EVM chains. Bulk files are priced per dataset.