Mirrored from filpgf.io — ProPGF Batch 3 (Karma program 1479, application 6a3174e913ef2aeed7ce0d8d, status: pending). Contact details redacted; canonical application lives on filpgf.io. 1.1 Project Name Filecoin SP Operations Intelligence 1.2 Project Github https://github.com/neti-software/llm-assistant-for-code-repos The project will be developed as a fully open-source, self-hosted solution. If the current proof-of-concept repository is private, we will either open-source it as part of Milestone 1 or create a new public repository for the production grant work. 1.4 Team Lead/Point of Contact Damian Pudło, Blockchain and Compliance Expert, preferred channel: Filecoin Slack @Damian Pudło or via e-mail 1.5 Category [ "RFP 2 - Storage provider economics and growth" ] 1.6 Open Source Status Fully Open Source 2.1 Project Summary **Project Description:** Operating a Filecoin Storage Provider requires expertise across multiple complex systems: Lotus, Curio, Boost, Singularity, Filecoin Pay, storage infrastructure, networking, sealing pipelines, deal onboarding, payment workflows, retrieval flows, and on-chain operations. In practice, Storage Provider issues rarely belong to only one component. A deal onboarding problem may involve Boost, Lotus, provider configuration, storage system state, retrieval availability, data preparation tooling, or payment setup. A sealing or storage issue may require understanding Curio, host-level resources, logs, hardware behavior, network configuration, and Filecoin chain state. This creates a significant operational burden for both new and existing Storage Providers. The proposed project is an open-source, self-hosted AI Operations Intelligence designed as a one-stop operational assistant for Filecoin Storage Providers. The goal is to reduce the expertise required to operate Filecoin infrastructure safely, profitably, and efficiently. The Operations Intelligence will combine Filecoin-specific knowledge with local environment awareness. It will help operators troubleshoot issues, validate configuration, monitor operational health, understand logs, prepare for paid-deal workflows, and follow recommended remediation steps. The project will not be limited to Curio. Curio is an important part of the Storage Provider stack, but a useful operational assistant must understand the broader environment. The intended scope includes Lotus, Curio, Boost, Singularity, Filecoin Pay, storage infrastructure, host-level diagnostics, logs, configuration files, and operational workflows around deals, data onboarding, and paid customer readiness. **Existing Proof-of-Concept** This project builds on an existing proof-of-concept that already demonstrates the feasibility of the approach. The current system is a microservice-based application composed of three loosely coupled containers communicating over HTTP. The presentation layer is a Streamlit-based chat interface. It connects to the agent core through a gateway exposing an OpenAI-compatible chat/completions endpoint. The gateway is protected with token-based authentication and preserves session state across restarts. The agent core is a self-hosted LLM runtime. It currently uses models accessed through OpenRouter, with gpt-5.3-codex as the primary model and claude-haiku-4.5 as a fallback model. The model layer is replaceable, and the system is designed so that model providers can be changed without redesigning the rest of the architecture. The agent uses two main sources of knowledge: - An indexed vector knowledge base. - Local repositories mounted as persistent read-only volumes. The vector database is served by a separate internal search service. It is populated with embeddings generated by …[truncated] 2.2 Who does this work support? [ "Storage Providers", "Network Infrastructure" ] 2.3 Total Funding Requested (USD) $190,000 2.4 Milestones & Budget [ { "title": "Milestone 1: Open-Source Foundation and Full Stack Knowledge Indexing", "description": "Estimated duration: Months 1–2\nDeliverables\nPublic GitHub repository for the grant-funded project.\nOpen-source foundation for the Storage Provider Operations Intelligence.\nSelf-hosted deployment package.\nArchitecture documentation.\nLocal deployment guide for Storage Providers.\nBaseline indexing pipeline for the broader Storage Provider stack.\nIndexed documentation and repositories for Lotus, Curio, Boost, Singularity, Filecoin Pay, and related Filecoin operational materials.\nDocumentation of the existing proof-of-concept architecture and retrieval pipeline.\nInitial security and privacy model for local deployment.\nBudget Justification\nThis milestone funds the open-source foundation, repository publication, deployment packaging, and expansion of the existing proof-of-concept from a Filecoin knowledge assistant into a broader Storage Provider stack knowledge base. The work includes engineering, documentation, indexing pipeline improvements, and public release preparation.", "dueDate": "2026-09-30", "fundingRequested": "$38,000", "completionCriteria": "Acceptance Criteria:\nA public GitHub repository is available before or at the start of the grant.\nA Storage Provider or reviewer can deploy the Operations Intelligence locally using documented steps.\nThe system can answer technical questions using indexed documentation and repository knowledge across Lotus, Curio, Boost, Singularity, Filecoin Pay, and related operational materials.\nThe repository includes installation documentation, architecture overview, contribution guidelines, and roadmap.\nSensitive local data handling is documented clearly." }, { "title": "Milestone 2: Local Environment Awareness for Lotus and Curio", "description": "Estimated duration: Months 2–3\n\n**Deliverables:**\nLocal context collection framework.\nRead-only Lotus environment awareness.\nRead-only Curio environment awareness.\nCollection of selected operational data such as configuration state, service status, logs, node status, miner state, and selected health indicators.\nEnvironment-aware question answering.\nOperator-facing interface for asking questions about the local setup.\nPlugin-style integration structure for future tools.\nBudget Justification\nThis milestone funds the first layer of local environment awareness. Lotus and Curio are core operational components for Storage Providers, and the work requires safe read-only collectors, local data handling, environment-specific retrieval, and careful security boundaries.", "dueDate": "2026-10-30", "fundingRequested": "$40,000", "completionCriteria": "**Acceptance Criteria:**\nThe Operations Intelligence can answer questions using both indexed Filecoin knowledge and local environment context.\nThe system can s …[truncated] 3.1 Impact pathway The impact pathway for Filecoin SP Operations Intelligence is: Output → We will deliver an open-source, self-hosted AI Operations Intelligence tool for Filecoin Storage Providers. The tool will combine Filecoin-specific knowledge with local environment awareness, indexing documentation and repositories across Lotus, Curio, Boost, Singularity, Filecoin Pay, and related SP workflows. It will provide diagnostics, log analysis, configuration validation, health checks, paid-deal readiness checks, and remediation recommendations. Outcome → Storage Providers will be able to diagnose operational issues faster, understand configuration problems more clearly, onboard data and deals with fewer blockers, and prepare their infrastructure for paid customer workflows. New providers will face a lower operational learning curve, while existing providers will spend less time on repeated troubleshooting and manual support loops. Core Filecoin engineering teams and maintainers should also see fewer repeated support questions because common issues will be captured in reusable diagnostic workflows and knowledge base entries. Impact → A healthier and more efficient Storage Provider base strengthens the supply side of the Filecoin network. By lowering operational overhead, improving reliability, and helping providers become ready for paid-deal workflows, the project indirectly supports growth in paid onchain deals, improves Storage Provider sustainability, and increases the network’s ability to serve real customers with production-grade storage services. In practical terms, the project turns fragmented operational knowledge into a reusable intelligence layer. That change helps Storage Providers move from “searching, asking, and manually debugging” toward “diagnosing, validating, and acting with evidence.” This reduces friction across the network and supports Filecoin’s broader objectives around paid usage, provider profitability, and flagship client adoption. 3.2 Verification metrics N/A for direct onchain metrics. This project does not directly create onchain transactions, storage deals, payments, or protocol-level state changes. Its impact is operational: it helps Storage Providers troubleshoot, validate configuration, run health checks, and prepare for paid-deal workflows. Indirectly, the project may support externally verifiable Filecoin metrics such as paid storage deals, storage volume onboarded, and deal activity of participating Storage Providers. These can be checked through public Filecoin chain data, explorers, provider-level deal records, and ecosystem dashboards. However, we will not claim direct attribution unless the metric can be clearly linked to participating providers and documented usage of the tool. 3.3 References David Gasquez, Michael Madoff, Amin Bensadok, Marta Piekarska- Greater, Paweł Burgchardt, Duncan Ross, 4.1 Monthly Operating Burn [ "$10-$100K (small team)" ] 4.2 What % of total team monthly burn depends on this grant? Answer: 100% As a software house, we delegate our devs to projects full time. Meaning that 100% of the grant will be covering the developer team delegated to this project. This does not represent the burn of the broader company/organization The estimated monthly cost of the dedicated project team is approximately $32,000 per month. The requested grant amount of $190,000 covers a six-month delivery period, including engineering, integrations, open-source release management, documentation, testing with Storage Providers, and community feedback cycles. 4.3 If this grant is not awarded, what happens? We will apply again next time, becouse we know that ecosystem need it. 4.4 Core Team Mateusz Motyl – Senior AI & Full-Stack Developer A blockchain and fintech expert who already intimately understands the ecosystem's needs and challenges, having successfully built an LLM dedicated specifically to the Filecoin network. Adrian Wachel – Senior Full-Stack Developer An experienced developer specializing in building robust, high-performance web applications and scalable system architectures for complex digital products. Łukasz Wal – DevOps Engineer A dedicated infrastructure specialist ensuring reliable, secure, and highly available deployments for seamless project execution. Damian Pudło – Senior Analyst & Compliance Specialist An analytical expert bringing invaluable, hands-on ecosystem knowledge and regulatory insight from his - 0.2 FTEactive, ongoing role in the Filecoin-based FIDL project. - 0.2 FTE Marek Foss – Senior Blockchain Architect A Web3 veteran with an extensive track record of designing secure, scalable, and decentralized infrastructure across multiple prominent blockchain protocols. **Team Synergy & Ecosystem Alignment (Neti):** **Team Background** Our team has been involved in the Filecoin ecosystem for approximately three years. During this time, we have worked with Filecoin-related infrastructure, tooling, and operational workflows. We have practical experience with the challenges faced by Storage Providers, including configuration complexity, sealing operations, Curio workflows, infrastructure maintenance, and the difficulty of troubleshooting issues across multiple components. We have already built a proof-of-concept AI assistant that continuously indexes Filecoin repositories, source code, and technical documentation. This existing foundation allows the assistant to answer Filecoin-specific technical questions and reason across documentation and source code. The grant will allow us to extend this foundation into a practical Storage Provider Operations Intelligence with local environment awareness, operational integrations, diagnostics, and eventually approved action execution. **Why This Team?** - We are well positioned to deliver this project because: - We already have a working proof-of-concept. - We have hands-on experience in the Filecoin ecosystem. - We understand Storage Provider operational pain points. - We can combine AI engineering with Filecoin-specific infrastructure knowledge. - We are committed to releasing the project as open-source and self-hosted. - We are focused on practical operational value rather than research-only output. This provides a highly stable, plug-and-play team structure requiring absolutely zero onboarding time to start delivering value to the Filecoin ecosystem. **GitHub:** github.com/neti-software github.com/fidlabs 4.5 Has your team received a ProPGF grant or funding from PLFIF before? [ "No" ] 5.1 Key risks & dependencies Risk 1: Local environment diversity Storage Provider environments differ significantly in hardware, configuration, deployment style, operating systems, monitoring setup, and operational practices. Mitigation: We will design the system around modular integrations and configurable collectors. The first release will focus on selected Curio, Lotus, and system-level signals. We will document supported environments clearly and expand coverage based on Storage Provider feedback. Risk 2: Sensitive operational data The Operations Intelligence may access logs, configuration files, and operational metadata that providers consider sensitive. Mitigation: The project will be self-hosted by default. We will implement a clear local data model, avoid unnecessary data collection, document what is accessed, and provide operator controls. Any future hosted or external model usage will be optional and explicit. Risk 3: Unsafe operational actions Automated actions in infrastructure environments can create operational risk if executed without sufficient controls. Mitigation: The first phases will be read-only. Operational actions will only be introduced later with explicit approval, limited action scope, audit logging, and the ability to disable execution entirely. Risk 4: AI hallucination or incorrect recommendations AI systems can produce incorrect or incomplete guidance, especially in complex infrastructure contexts. Mitigation: Recommendations will be grounded in retrieved documentation, source code, local logs, and configuration evidence. Diagnostic outputs will include supporting context. The system will distinguish between confirmed findings, likely causes, and suggested next steps. Critical actions will require operator confirmation. Risk 5: Broad Storage Provider Stack Scope The Filecoin Storage Provider stack is broad and includes Lotus, Curio, Boost, Singularity, Filecoin Pay, host-level services, storage infrastructure, networking, deal workflows, and payment-related tooling. Attempting to support every component equally could create scope risk. Mitigation: The grant scope is phased by component and operational priority. Milestone 1 focuses on public release and broad knowledge indexing. Milestone 2 focuses on Lotus and Curio environment awareness. Milestone 3 expands into Boost and Singularity for deal and data onboarding diagnostics. Milestone 4 adds comprehensive health checks and Filecoin Pay awareness, with Filecoin Pay marked as a stretch area if full integration is not feasible. Milestone 5 focuses on safe actions, testing, documentation, and public release. This phased approach allows the project to present a credible one-stop-shop vision while keeping milestone delivery realistic. Risk 6: Adoption and feedback The tool must be tested by real operators to become useful. Mitigation: We will involve Storage Providers during …[truncated] Anything else you want to share that we didn't ask? **Additional Context** This project is designed as practical public-good infrastructure for the Filecoin Storage Provider community. The current proof-of-concept already demonstrates that an AI assistant can index Filecoin documentation, repositories, and source code and provide useful answers about Filecoin architecture and operations. The proposed grant work moves the project from knowledge assistant to environment-aware operations Operations Intelligence. The long-term vision is a self-hosted operational layer that helps Storage Providers: - Understand their infrastructure. - Detect issues earlier. - Troubleshoot faster. - Validate configuration. - Learn from community operational knowledge. - Execute safe maintenance workflows with explicit approval. - Reduce reliance on a small number of highly experienced operators. By making Storage Provider operations easier and more repeatable, the project can help improve provider retention, reduce operational costs, support paid-deal readiness, and strengthen the long-term resilience of the Filecoin network. In summary, the project will deliver an AI operations agent for Filecoin Storage Providers that helps automate day-to-day operational work, simplifies onboarding of new datasets and paid-deal workflows, and supports ongoing infrastructure maintenance. By reducing the manual expertise required to run and troubleshoot Filecoin infrastructure, the Operations Intelligence lowers the barrier to entry for new Storage Providers and helps existing operators run their infrastructure more efficiently and sustainably. **Expected Impact** The expected impact of this project is to reduce operational complexity across the Storage Provider ecosystem. In the short term, the project will provide an open-source tool that operators can deploy locally to ask environment-aware questions, inspect operational state, and receive troubleshooting recommendations. In the medium term, the project will create a reusable diagnostics and operational knowledge framework for common Filecoin Storage Provider issues. In the long term, the project can become an extensible operational Operations Intelligence for Filecoin infrastructure, supporting additional integrations, safer automation, and improved operational standards across the ecosystem. **Sustainability After the Grant** After the grant, the project can continue as an open-source tool maintained by our team and the Filecoin Storage Provider community. Future sustainability options may include: Community contributions. Additional ProPGF or RetroPGF funding based on demonstrated impact. Paid support or implementation services for Storage Providers who want help deploying or customizing the tool. Integration work with Filecoin ecosystem teams and tooling providers. Expanded support for paid-deal workflows and production-readine …[truncated] Contributing to Core Infrastructure? Filecoin Storage Provider SP Operations Intelligence will contribute open-source operational infrastructure for Storage Providers. The project will provide a self-hosted AI assistant that combines Filecoin-specific knowledge with real-time awareness of the operator’s local environment. Storage Providers depend on many complex systems, including Curio, Lotus, storage infrastructure, sealing pipelines, network configuration, deal management, and on-chain operations. Today, much of the practical operational knowledge required to run these systems is concentrated among experienced operators and scattered across documentation, repositories, issue threads, community discussions, and internal runbooks. The Operations Intelligence will turn this knowledge into an accessible, environment-aware operational assistant. It will help providers understand their configuration, diagnose common problems, validate operational state, interpret logs, and receive concrete remediation recommendations. Over time, it will also support approved operational actions through integrated APIs. Objective 1 Indirect Objective 2 Direct Objective 3 Direct Open Source Context NA