Application for Filecoin ProPGF Batch 3

Goldsky

https://github.com/goldsky-io/

https://goldsky.com/

Jorge Bours, Partnerships, Slack

Goldsky maintains production indexing infrastructure for Filecoin and FEVM, including subgraph hosting and real-time data pipelines used by applications, ecosystem teams, and analysts who need reliable access to onchain data without running their own archive nodes or indexing stack. Filecoin's developer documentation lists Goldsky as a supported FEVM indexer for Filecoin mainnet and testnet, and Goldsky provides both no-code and CLI paths for deploying subgraphs. This infrastructure is critical because FEVM is not a generic EVM environment from an indexing perspective. Goldsky has absorbed Filecoin-specific graph node, RPC, and operational complexity at the infrastructure layer so each application team does not need to debug and maintain its own bespoke indexing path. That work reduces the cost of building on Filecoin, keeps existing applications operational, and gives new builders a familiar GraphQL/subgraph interface for smart-contract data. Current and intended dependents include Sushi, Secured Finance, FilOz, FIL Builders, FilBeam, Filecoin Onchain Cloud, ecosystem builders onboarded through Filecoin developer relations, and Open Source Observer. Goldsky also provides dedicated Slack-based support and technical assistance to Filecoin teams and builders when they deploy, migrate, backfill, or debug subgraphs and data pipelines.

Partial

Goldsky is a partial open source project. The managed cloud service remains proprietary because it funds the engineering, operations, reliability, and support work required to run production-grade indexing infrastructure for many chains, including Filecoin. However, a substantial portion of the non-open source components have no relevance to the broader ecosystem; they include billing infrastructure, RBAC, and other “hosted-solution-specific” infrastructure. The core components of value of both indexing frameworks: Turbo and Subgraphs, are open source. This model is aligned with Filecoin's 2026 objectives because it keeps critical ecosystem infrastructure available for self-hosting, while reducing duplicated infrastructure spend for builders who would otherwise need to run their own indexers and make duplicative RPC calls.

Goldsky will continue its existing relationship with Filecoin by sustaining the indexing infrastructure and ecosystem support that Filecoin and FEVM teams use to build applications, run analytics, and monitor onchain activity. The base scope continues production subgraph hosting and technical support for active ecosystem applications such as Sushi, Secured Finance, Filecoin Onchain Cloud, and other Filecoin builders, while adding Turbo (powered by the open source Streamling project) as a new Filecoin data infrastructure workstream. Turbo expands what Goldsky can support beyond subgraph-style access by enabling faster and more flexible real-time data pipelines for Filecoin EVM workflows. One of the key advantages of this is to enable different access patterns (ie. large-scale analytics access via OLAP databases like ClickHouse, and AI-focused workloads via vector databases like TurboPuffer).

$86,000

N/A

Indirect

Indirect

**Output**: Goldsky maintains production Filecoin/FEVM subgraph infrastructure, supports active ecosystem builders, adds Turbo support for Filecoin EVM data workflows. **Outcome**: Existing Filecoin applications operate with reliable indexed data, builders gain a broader Turbo-backed data path without running custom indexers, and analysts gain a shared public dataset instead of rebuilding chain ingestion independently. Filecoin Onchain Cloud and other ecosystem teams also gain a more dependable data layer for monitoring application activity, payments, contract interactions, and developer adoption. **Impact**: This work supports Filecoin's 2026 objective to drive network profitability and cryptoeconomic credibility by giving ecosystem teams better public data for measuring real adoption, rather than relying on anecdotal and/or self-serving reporting. It supports flagship client adoption by lowering the integration burden for teams with persistent storage, compute, and onchain data needs, and enables classes of data workloads that are not otherwise possible. In short, Goldsky maintains the data infrastructure that lets Filecoin applications, customer workflows, and ecosystem measurement systems function effectively and grow.

Goldsky is infrastructure that enables applications and measurement systems; primarily operating read infrastructure in the scope of work proposed here: https://goldsky.notion.site/SHARED-Filecoin-ProPGF-Batch-3-Application-Section-3-2-Verification-Metrics-38195cab936e80139bd1e3f88f5cce58 While the verification metrics specify onchain-only, we still have included the above for review and discussion; our scope of work primarily relates to pushing onchain data to offchain datastores and thus it’s difficult to measure success purely from onchain data. However, we want to ensure we are held accountable to delivering value to the Filecoin ecosystem and thus propose some alternative metrics here instead.

We can make introductions for reference calls to any of the teams that are currently or plan to use Goldsky for accessing Filecoin data, incl. FilOz, SushiSwap, and/or OSO. Since these applications are publicly visible we have not named specific individuals here out of respect for their privacy, but would be happy to do so in private communications.

- **Jorge Bours**, Partnerships, key POC for this proposal application. - Based in Denver, CO, Jorge leads partnerships at Goldsky and serves as the primary liaison between Goldsky and the ecosystem and protocol teams it supports. He owns the Filecoin relationship end to end - coordinating onboarding, enablement, and ongoing technical support for ecosystem builders deploying, migrating, or debugging subgraphs and data pipelines, and aligning Goldsky's infrastructure roadmap (subgraph hosting and Turbo) with the needs of Filecoin/FEVM teams. He works closely with the engineering and technical services teams internally to translate partner requirements into delivered infrastructure, and is the main conduit for reference introductions, milestone reporting, and ecosystem coordination throughout the grant. - **Jeff Ling**, cofounder and CTO, primary owner of subgraph and pipeline implementation - Based in Salt Spring Island, BC, Canada, Jeff is a Goldsky co-founder. He worked on several startups before this and held management roles at a couple of companies. His stated focus is sharpening Goldsky's capabilities by balancing the team's existing state, its capabilities, and business goals - and he places a high value on getting things done as simply and straightforwardly as possible. As CTO he coordinates engineering and product, but his role stretches well beyond that: works closely with ingestion team, helps with technical customer support, lead the new streaming engine revamp, curates new data sets for chains, works with customers to iterate on new use cases and is a resource to all of our partners. At the time of his bio he was also acting PM. Acts as head of engineering. Prev. experience includes VP of Engineering at TrueAccord (TrueML), CTO at Uppercase, and an early employee at Heap. - **Hakon Amdal**, ingestion team lead - Based in Stavanger, Norway, Håkon joined Goldsky in July 2024. He holds a Master's in Computer Science from NTNU and built his career in data engineering -starting at Schibsted, then joining Dune to build out their data platform, Arrakis, where he moved from decoding pipelines into an engineering manager role. He co-leads "All Things Ingestion" alongside Sek Fook - both leading/structuring the ingestion process and staying hands-on scaling the infrastructure. He owns data quality, and his deepest expertise is at the intersection of data and crypto, particularly EVM and Solana decoding. - **Sek Fook Tan**, ingestion team - Based in Kuala Lumpur, Malaysia, Sek Fook is one of the longer-tenured engineers at Goldsky - joined September 2022. Originally from Melaka, he did a Mechanical Engineering degree in Johor Bahru and came into software through math and machine learning research before going down the web3 rabbit hole. At Goldsky he works on blockchain data ingestion and infrastructure - the same ingestion domain Håkon co-leads, so the two are close collaborators. He's drawn to systems programming (OS, databases), DeFi, and next-gen blockchain data infra. - **Kyle Quintal**, ingestion team - Based in Montreal, Canada - joined late July 2025. He's a software engineer by training who's been in the startup ecosystem since 2018, with an MSc in AI + Cybersecurity and a handful of published academic papers. His roles have been mostly technical, though he likes getting involved in the business day-to-day while keeping his focus on engineering. Kyle joined Goldsky’s ingestion team and specialized in both EVM and non-EVM chain ingestion and decoding. - **John Clover**, technical services - Based in Raleigh, NC. At Goldsky he runs the Technical Services team, which covers most customer-facing and technical functions - Support, Customer Success, and Services. So he's your point person for anything touching customer technical support, troubleshooting, pipeline/subgraph config, RPC debugging and anything else you can think of. - **Jonathan Archer**, technical services - Based in San Francisco and joined the team earlier this year. At Goldsky he's a solutions engineer on John Clover's Technical Services team, working in generalist mode supporting customers and prospects. Prior companies include TPL, ITVS, Large, Heap, and Sigma. Works day to day with our customers and enabling product use cases.

Key dependencies are access to reliable Filecoin/FEVM RPC or archive data, and agreement on the exact public archive schema. Technical risks include FEVM-specific indexing behavior, backfill complexity, data freshness during high-volume periods, and the amount of historical data required if traces or decoded events are included. Delivery risk is manageable if scope is fixed early around specific tables, block ranges, freshness targets, and named verifier artifacts. Our existing experience on the subgraph indexing scope significantly de-risks this, and we have pre-discussed the target schema for public datasets with key users (eg. OSO). There are limited team risks; there are multiple members on the Goldsky team familiar with our relevant codebase and the Filecoin institutional knowledge from our subgraph indexing scope of work.

Goldsky's request is intentionally below the Batch 3 soft cap and materially below the program's stated average grant size, while still addressing a core infrastructure gap. The proposal focuses on continuing an existing Filecoin data infrastructure relationship, adding Turbo support, and optionally leaving Filecoin with a durable public FEVM archive dataset. This additional milestone is optional, we have seen a high degree of demand from this and usage of similar datasets on other networks has been strong, but if the committee does not feel it is necessary at this time, we are happy to remove it from consideration.

None!