Hi everyone,
TLDR:
I’ve spent 4+ years building in the Polkadot ecosystem and a lot of that time was spent helping new-comers navigate ecosystem nuances, data and practices.
So I built Blockpedia, a tool that lets you ask questions about the Polkadot ecosystem in plain English and get answers with verifiable citations (no AI guess-work).
Built for Polkadot’s “product era beyond crypto-natives ”, the beta is live at blockpedia.ai
Just want feedback from people who use this data and people who are still figuring the ecosystem out.
What it does
You type a question. Blockpedia figures out which sources to hit, fetches the data, and gives you an answer with citations you can click to see the raw source. You see every step of the process, what and why each decision is made via the "Thinking " tab and verify the data via the "Provenance " tab.
A few basic examples of what works today:
- “What is referendum 1839 about and who voted on it?”
- “What’s Saxemberg’s address on Polkadot?”
- “How much DOT was transferred on Polkadot this week?”
- “How does OpenGov governance work?”
- “What are people saying about Agile Coretime?”
- “What’s the current DOT price and how much has the treasury spent?”
- “Is the marketing bounty still going on ?”
- “How much has project-XYZ been granted from the treasury until now?”
The idea is that whether it’s someone who just heard about Polkadot through the TDOT ETF, or even an experienced delegate evaluating a proposal, they can type a question the way they’d ask a colleague and get a real answer backed by real data.
Why this exists
Gavin said in his 2025 roundup that Polkadot’s focus is shifting from technology and protocol towards platform and product, with a mission to reach “the population beyond crypto-natives.” W3F is winding down centralised support in favour of “systems and products that make support less necessary.” The ecosystem is entering its product era.
But right now, if someone new comes to Polkadot and wants to understand what’s going on, what a referendum is about, who voted on it, how much they’re asking USD, they need to know how to jump between Subscan, SubSquare, the wiki, CoinGecko, then maybe the Forum. Five tools, five interfaces, five mental models. If you’ve been here for years you know which tool has what. If you haven’t, you’re already lost.
The data exists. The teams maintaining these tools do amazing work. But getting a full picture means knowing which tool has which data and manually stitching it together, and that’s a barrier that hits newcomers hardest.
Why not just use ChatGPT?
Fair question. One obvious answer is that ChatGPT doesn’t have access to live Polkadot data. But even if it did, there’s deeper problems.
On the infrastructure side, Blockpedia maintains its own Subsquid indexer for on-chain analytics, connects directly to the chain via a Smoldot light client for real-time lookups, and integrates with 9 ecosystem leading data sources, each with its own query logic, caching, and error handling. When one source is down, the system replans and tries alternative sources automatically. That kind of reliability doesn’t come from plugging an API key into a prompt, it comes from expertise and active maintenance in a rapidly developing ecosystem.
Moreover, Polkadot has its own vocabulary. Conviction multipliers, origin tracks, fellowship ranks, the difference between a treasury proposal and a treasury spend, which governance data lives on SubSquare vs Polkassembly vs on-chain. Getting useful answers requires knowing these distinctions. I’ve spent years building tooling in this ecosystem and that domain knowledge is baked into how Blockpedia interprets questions, picks the right source, and structures its answers so it’s not a generic AI wrapper.
And most importantly, with Blockpedia, the LLM is just a language layer, not the source of truth. Every claim in a response has an inline citation pointing to the exact source and raw data. If the data doesn’t exist or the sources don’t have it, Blockpedia says so instead of guessing.
How it works
When you ask a question, Blockpedia first analyses what you’re asking and breaks it into steps if needed. You can watch this happen in real time through the “Thinking” tab.
For each step, it searches across 9 data sources (Subscan, SubSquare, Polkassembly, Polkadot Forum, CoinGecko, CryptoCompare, the Polkadot Wiki, a custom Subsquid indexer, and PAPI with Smoldot for direct chain queries), picks the right one, generates the query, and fetches the data. Steps that don’t depend on each other run in parallel. If something fails or comes back incomplete, it replans and tries a different source.
You get an answer with inline clickable citations which link to the raw data in the “Provenance” tab so you can verify exactly where every claim comes from. It comes with ecosystem best-practices, for example, on-chain addresses in the response show identicons and resolve to on-chain identities via the People Chain(s) when available.
If your question is ambiguous, it asks you to clarify before fetching anything instead of guessing.
Blockpedia doesn’t replace Subscan or SubSquare or any of the tools the ecosystem already has. It sits on top of them on a single conversational interface. The work these teams have done is what makes this possible. This is additive.
Networks supported: Polkadot, Kusama (and their Asset Hub).
Built with LangGraph, Gemini, Next.js, PostgreSQL, PAPI, Subsquid
Where it is right now
Governance. You can ask about referenda, proposals, bounties, treasury spending, delegate activity, and voting records across Polkadot and Kusama. It pulls from both Polkassembly and SubSquare, so you get coverage even if one source is missing something.
On-chain data. Account lookups, validator info, commission history, staking details, transfers, and decoded extrinsics via Subscan. On-chain identity resolves directly from the People Chain through a Smoldot light client, no RPC provider involved.
Market data. Current prices, market caps, trading volume, all-time highs, and historical price data for DOT and KSM. Uses both CoinGecko and CryptoCompare so there’s no gap in historical depth.
Analytics. Transfer volume trends, top holders, account distribution, daily activity breakdowns, and Asset Hub token tracking (USDT, USDC). This runs against a custom Subsquid indexer I maintain.
Knowledge. Questions about how Polkadot works (OpenGov, XCM, Agile Coretime, staking mechanics, parachains) pull from the Polkadot Wiki. Useful for people still learning the ecosystem.
Community. It searches Polkadot Forum discussions, so you can ask what people are saying about a specific topic or proposal.
In progress
-
Expanding parachain-specific coverage beyond Polkadot and Kusama relay chains
-
More big data querying and analytics capabilities
-
Faster responses on complex multi-step queries
-
Charts, graphs, and data visualisations in responses
Please try it
I’d like to hear from anyone interested, be it people deep in governance who can tell me if this helps them with queries they deal with day to day or people newer to the ecosystem who can tell me if this makes things less confusing.
I’ll be in the replies.
I truly appreciate you taking the time to look at this.