We’re opening this thread to share Polkassembly’s upcoming Treasury proposal for governance feature development, shared AI tooling and infra upgrades, and to gather feedback before it goes on-chain.

Over the last year, Polkassembly has continued to be the main coordination layer for OpenGov. In this period we’ve gone well beyond “keeping the site running” and have shipped a lot of new work around AI-native governance tools, treasury analytics, identity/relationships, and a modern, cheaper infra stack. This proposal requests $224,000 USDT, the majority of which is retroactive funding for work already delivered or close to completion, with the rest covering final milestones through December 2025.

1. Governance transparency & treasury analytics

We’ve been turning Polkassembly into more than a proposal list by adding proper analytics and reporting on top of OpenGov data. This includes:

• A Judgement Explorer that surfaces registrar activity, fees and turnaround times in one place.
• Treasury dashboards tracking balance, inflows/outflows, liabilities and spend trends, along with clearer breakdowns by beneficiary and category.
• Early treasury reporting and AI-based spend insights so that recurring reports and categorisation can be generated consistently, rather than manually.
• Richer delegation, bounty and ROI analytics that make it easier to see how funds and voting power are actually being used.

All of this is being built as reusable public infra (APIs, components and datasets) so other ecosystem teams can plug it into their own products instead of rebuilding the same plumbing.

2. Identity, proxies and relationships

We’re also investing in the “who is who” layer of governance:

• Cleaner sub-identity flows so that setting up identity hierarchies is less confusing.
• Account relationship and proxy mapping to make it obvious who is acting on whose behalf.
• A unified Proxy Explorer for governance, staking and identity proxies.

The goal is to make on-chain roles, responsibilities and links between accounts much more transparent, in line with the expectations set by the Social Contract.

3. Klara and the shared AI layer

A large part of this proposal is about Klara, our governance-focused AI layer built on top of Polkassembly and on-chain data.

Klara is not a generic chatbot. It is backed by:

• A curated governance knowledge base that combines Polkassembly discussions, OpenGov events, forum posts and official docs in a machine-readable way.
• A decentralised knowledge layer (DKG on OriginTrail) so knowledge assets are verifiable and can be extended by other teams.
• A chat interface and SDK for reading proposals, analysing arguments, comparing options and guiding users through actions.
• DelegateX, an intelligent delegation engine on CyberGov where users define profiles and criteria, and the AI delegate prepares or executes votes with clear rationale.

We’re designing this as shared ecosystem infrastructure: wallets, parachains and other governance tools should be able to use the same AI base layer, rather than every team spinning up isolated assistants.

4. Asset Hub migration, multi-chain indexing & infra revamp

A lot of the work we’re asking retroactive funding for is infra-heavy and not directly visible on the UI, but it is what keeps governance usable as Polkadot evolves:

• A unified, multi-chain indexer architecture that handles Relay Chain and Asset Hub data in real time while preserving full governance history.
• A refactored Next.js frontend, APIs, caching and database schema that are chain-aware but still feel consistent for users.
• Migration of our infra to GCP with better CI/CD, security, observability and scaling, and integration of PAPI for more stable multi-chain RPC.

This isn’t just maintenance. These changes have already:

• Reduced page load times by roughly 60%.
• Improved API response times by around 35%.
• The GCP migration and other infrastructural updates have helped us optimise and reduce maintenance costs. These optimisations will save us, and hence the Treasury, over $150k in infrastructure and maintenance costs next year, and over $400k over the next three years.

On top of that, we’re packaging common flows (wallet connect, proposal lists, vote tracking, delegation and analytics widgets) as npm packages, APIs and UI kits so other builders can reuse them.

5. Scope, alignment and value

We’ve scoped this proposal carefully around the 2025 Social Contract and recent community feedback:

• Fellowship workflow work has been removed and will be handled separately, so this request focuses on AI, analytics, identity/relationships and infra upgrades.
• There is no overlap with Ref. 1463’s maintenance scope; we are not asking for the same work twice.
• The AI work is scoped as open and reusable infrastructure (SDKs, knowledge assets, vector indexes), not a closed product.

Overall, this is a request to:

• Retroactively fund deep, already-delivered engineering on multi-chain governance infra and cost optimisation.
• Support AI and analytics capabilities that serve the wider ecosystem, not just Polkassembly.
• Reduce long-term Treasury spend by avoiding duplicated infra and making our own infra materially more efficient.

  
  Next steps

  We’re sharing this draft here to let everyone review the scope, numbers and priorities before the proposal goes on-chain.Please leave any questions, concerns or suggestions in the comments. We’re happy to go deeper on specific parts (AI layer, treasury analytics, infra cost model, Asset Hub migration, etc.)

Here's the detailed discussion document for review.