Overview

This proposal asks the Treasury to support core governance and infrastructure work that makes Polkadot governance and treasury activity easier to understand and analyse, and that establishes Klara as a shared AI layer for the ecosystem. Together, these changes create a durable base that wallets, chains and applications can reuse instead of each rebuilding their own governance, data and AI pipelines.
Over the last year, the Polkassembly team has focused on four main areas:

1. Governance transparency and treasury analytics
2. Identity, proxies and account relationships
3. Klara as a shared AI layer for Polkadot
4. Asset Hub migration, multi-chain indexing and infrastructure optimisation

The request is for 224,000 USDT, of which around 90% is retroactive for work that is already live or functionally complete, with the remaining milestones landing by December 2025. Fellowship workflow work has been deliberately separated out and is not part of this request.

Polkassembly remains the primary place where Polkadot governance is read and discussed, handling around 80,000+ monthly visitors, 93.69% of discussions and 82.33% of comments on Polkadot governance today.

1. Governance transparency and treasury analytics

Polkadot’s governance and treasury history now spans multiple tracks, sub-treasuries and chains. A core aim of this work is to give delegates, reviewers and community members a single, reliable view of what is happening, without needing spreadsheets and ad-hoc tools.

Work delivered or underway includes:

• A Judgement Explorer that tracks judgement requests, registrars, fees and turnaround times in one place.
• Treasury dashboards that show balances, historical spending, liabilities and basic efficiency trends, including coretime- and spend-related views.
• Treasury analytics that track balance, inflow, outflow and liabilities, with AI-assisted spend categorisation and report generation aligned with existing community reporting formats.
• Delegation statistics that show delegation flows and performance.
• Bounty analytics that surface completion rates, curator performance and category-level effectiveness.

These are being built as public infrastructure: APIs, queries and components that other teams can plug into their own products rather than recreating the same pipelines.

Planned milestones in this area include:

• Extending treasury views with multi-cycle comparisons, recurring beneficiaries and under-used tracks.
• Hardening APIs and export formats so wallets, explorers and researchers can pull the same views directly.
• Adding simple status and impact fields so long-term outcomes of treasury spending are easier to follow.

2. Identity, proxies and account relationships

As Polkadot moves toward more application-driven use and individuality / PoP-aligned systems, it becomes more important to see who is acting and how accounts relate to each other across governance, staking and identity.

Delivered or in progress:

• A simplified sub-identity flow to make identity hierarchies easier to create and manage.
• Account relationship mapping that visualises how accounts and proxies relate.
• A unified Proxy Explorer that lets users create and manage governance, staking and identity proxies from a single interface.

This responds directly to long-standing user and delegate feedback: people want to understand which entities sit behind which accounts and actions, especially as PoP and individuality progress.

Planned next steps:

• Extending relationship views to cover Fellowship, collectives and other system roles where this is meaningful.
• Exposing relationship data through simple APIs so other tools do not each implement their own custom logic.
• Improving explanations and examples, so users can understand common patterns without needing to inspect raw chain data.

3. Klara as a shared AI layer for Polkadot

Klara is an AI layer for the Polkadot ecosystem. It has been deployed and tested in governance first on top of Polkassembly because that is where most governance data and activity already are, but it is not limited to governance or to Polkassembly.

Klara today includes:

• A curated governance knowledge base built from Polkassembly archives, on-chain data, OpenGov activity and core documentation.
• A decentralised knowledge layer using OriginTrail DKG, so knowledge assets are verifiable, traceable and can be extended by other teams.
• A chat interface and SDK that helps users understand proposals, compare options and navigate governance actions.
• DelegateX on CyberGov: an intelligent delegation engine where users define profiles and evaluation criteria, and an AI delegate prepares or executes votes with clear rationale.

In the first weeks after launch, Klara handled well over 1,200 queries in 15 days, and more than 2,500 questions in the first month, mostly around bounties, sub-treasuries, tracks, conviction and referendum effects. This confirms that even with fewer treasury proposals, questions and complexity for users have not decreased.

Klara is designed as shared infrastructure, not a one-off chatbot:

• Teams can plug their own data into Klara’s index to specialise answers for their chain or application.
• Chains can use Klara as a chain-specific assistant while benefiting from a common Polkadot knowledge base.
• Applications can integrate Klara via SDKs or APIs to power contextual help, reviewers, co-pilots or AI delegates, instead of building separate AI stacks from scratch.

From the Treasury’s perspective, this helps avoid duplicated AI builds. A serious AI governance assistant typically costs at least 50,000 USD for a single team to implement. If multiple ecosystem teams leverage Klara instead of rebuilding their own pipelines, that avoids hundreds of thousands of dollars in duplicated costs.

Planned milestones for Klara:

• Finalising and documenting SDKs and APIs so external teams can integrate without heavy hand-holding.
• Extending Klara to more non-governance domains such as developer documentation, XCM routing, PoP / individuality and identity flows.
• Providing reference integrations (for example a DeFi helper or pallet co-pilot) that teams can fork rather than starting from zero.

4. Asset Hub migration, multi-chain indexing and infrastructure

The migration of governance and other pallets to Asset Hub has been one of the most significant infrastructure changes in Polkadot’s history. Supporting this required a full rethink of how Polkassembly indexes and serves governance data.

Key outcomes delivered:

• A unified indexer architecture capable of handling Relay Chain and Asset Hub governance in real time while preserving historical continuity.
• A refactored Next.js frontend, APIs, caching layer and database schema that are chain-aware but still coherent for users.
• A full migration to Google Cloud Platform (GCP), including CI/CD, security hardening, observability and integration with PAPI for more predictable multi-chain RPC.

The measured impact so far:

• Page load times reduced by around 60%.
• API response times reduced by around 35%.
• Maintenance and infra optimisation projected to save over 150,000 USD in the next year and over 400,000 USD over the next three years, compared with the previous setup.

In parallel, Polkassembly is building developer-facing infrastructure:

• Public APIs across multiple networks and open-source libraries and components.
• A growing set of npm packages and UI kits for common governance flows such as wallet-connect, proposal lists, vote tracking, delegation and analytics widgets.
• An open usage guide for builders to adopt these APIs and packages quickly.

This positions Polkassembly as long-term governance infrastructure rather than just a front-end.

Remaining milestones in this area:

• Final packaging and documentation of reusable components for external teams.
• Continued tuning of infra and cost controls so running the platform stays reliable and cost-efficient as more chains and features are added.

5. Cost, value and alignment with the Social Contract

The total request of 224,000 USDT is primarily retroactive and covers a year of work on:

• Multi-chain governance infrastructure and GCP migration
• Treasury analytics and reporting
• Identity and proxy tooling
• Klara and DelegateX as a shared AI layer

The internal value-versus-cost analysis breaks this into four main cost blocks (infrastructure and GCP, analytics, Klara / DelegateX, identity and proxies) and compares each to conservative estimates of value to the ecosystem. Even on cautious assumptions and only counting features already delivered, the work is expected to generate:

• Over 400,000 USD in direct infra savings over three years
• High six to low seven figures in better-aligned Treasury allocation and avoided waste
• Hundreds of thousands of dollars in avoided duplicated ecosystem builds

This proposal is structured to align with the 2025 Social Contract (Ref. 1463) and current guidance:

• The focus is on infrastructure and shared layers: governance data, analytics, identity and relationships, AI infrastructure and indexing.
• Fellowship workflow work is explicitly excluded from this scope and handled separately.
• There is no overlap with existing maintenance referenda; this proposal does not bill twice for the same work.
• Most of the work is delivered; remaining milestones are specific, time-bounded and publicly reportable.

The Polkassembly team has been focused solely on Polkadot governance for more than four years and continues to run a full-time team in an environment where many ecosystem teams are dealing with uncertainty about future structure and support. This proposal aims to keep that continuity, narrow the scope to core infra, reduce long-term costs and ensure more of what we build can be reused by other teams across the network.

Detailed Proposal Doc.
FAQs based on feedback from community members and discussion with Parity & Foundation.

We look forward to your support and ideas on expanding our AI work and helping move Polkadot & Polkassembly into its app-centric era.

Regards,
Polkassembly Team