// devcon 8 mumbai · oct 31 – nov 8, 2026

IPFS Planet Expedition

Resilient Storage @ Devcon’s Bunkers

Build with IPFS. Connect with libp2p. Make decentralized storage work in the real world.

Quest rewards

Top 3 quest winners receive free Devcon 8 Mumbai tickets

Build an original contribution, demonstrate it with evidence, and help move decentralized storage forward. Working software, upstream pull requests, cross-implementation tests, research, benchmarks, and complete application demonstrations are welcome.

Evaluation criteria

From Go-based Kubo and Boxo to JavaScript-based Helia and Python-based libp2p applications, participants will learn how to build, test, and scale decentralized systems across languages and deployment environments.

A special focus of this expedition is SocialCalc and EtherCalc: collaborative spreadsheets and offline-first applications that can serve as practical testbeds for content-addressed storage, replicated data, healthcare workflows, and AI-assisted document automation.

Rather than stopping at tutorials, the expedition turns real ecosystem challenges into quests with concrete deliverables, reproducible experiments, working demonstrations, and opportunities to contribute upstream.

// the mission

Make decentralized storage resilient, composable, and accessible

IPFS provides a foundation for content-addressed data and peer-to-peer retrieval. The challenge is turning those capabilities into dependable systems that developers and organizations can deploy, operate, and maintain.

We invite participants to explore:

Storage and replication: Kubo, Boxo, IPFS Cluster, content addressing, pinning, and retrieval across multiple nodes.
Cross-language development: Go, JavaScript, and Python libraries, interoperability, and shared testing infrastructure.
Offline-first applications: Local-first data, intermittent connectivity, portable IPFS deployments, and recovery after network outages.
SocialCalc and EtherCalc: Collaborative spreadsheets, healthcare applications, AI agents, MCP tools, and IPFS-backed document workflows.
Gateway resilience and policy: Self-hosted gateways, AutoTLS, malicious-content awareness, and reduced dependence on public infrastructure.
Distributed AI and data: Content-addressed models, datasets, embeddings, execution artifacts, and machine-to-machine workflows.
Privacy and verifiability: Encryption, content integrity, provenance, and reliable access to distributed application data.

The goal is to make IPFS easier to use not only for infrastructure experts, but also for application developers, students, healthcare innovators, AI builders, and small teams operating their own infrastructure.

// quest board

Eight quests. One resilient storage ecosystem.

8 of 8 quests

QUEST 01 · Foundations and interoperability

Beginner–Intermediate

Learn IPFS Across Go, JavaScript, and Python

Explore how IPFS applications use content identifiers (CIDs), blocks, UnixFS, peer discovery, transport connectivity, and content retrieval.

Challenge, deliverables & references

Build a small application that stores and retrieves content using a selected IPFS implementation. Compare Kubo and Boxo in Go, Helia in JavaScript, and relevant Python networking or IPFS integration options. Identify which capabilities are implemented directly by each library and which depend on additional components.

Investigate how libp2p enables peer connectivity, and how differences in implementation, configuration, and deployment affect retrieval.

Deliverables

A working IPFS application with reproducible setup instructions.
An architecture diagram showing the storage and networking layers.
A comparison of relevant Go, JavaScript, and Python libraries.
Retrieval tests and documented limitations.
At least one improvement proposal, issue, test, or upstream contribution.

Explore

QUEST 02 · Resilient storage infrastructure

Intermediate–Advanced

Replicate, Retrieve, and Recover with Kubo, Boxo, and IPFS Cluster

What happens when an IPFS node disappears, a storage machine fails, or a network becomes temporarily unavailable?

Challenge, deliverables & references

Build a multi-node storage environment using Kubo, explore the underlying Boxo libraries, and use IPFS Cluster to coordinate pinning and replication across machines.

Measure the relationship between replication factor, storage overhead, retrieval performance, and recovery after node failure. Investigate whether content remains retrievable from surviving nodes and how the system behaves when connectivity returns.

Experiments

Replicate content across multiple independent nodes.
Compare retrieval from local, nearby, and remote peers.
Stop nodes and observe content availability and recovery.
Test re-pinning, cluster coordination, and replication behavior.
Measure storage overhead, latency, and recovery time.

Deliverables

A working multi-node deployment, reproducible test scripts, a replication and retrieval benchmark report, and a documented failure-recovery strategy.

Stretch goal: Contribute an automated resilience test or reusable benchmark to the IPFS ecosystem.

Explore

QUEST 03 · Gateway resilience and trust

Intermediate–Advanced

Build Policy-Aware, Self-Hosted IPFS Gateways

As developers operate more of their own IPFS infrastructure, how can self-hosted Kubo gateways make informed decisions about potentially malicious or problematic content?

Challenge, deliverables & references

Explore the relationship between content retrieval, gateway policy, content identifiers, and the existing Bad Bits infrastructure.

Prototype an AI-assisted gateway policy layer that uses a retrieval-augmented generation (RAG) system over historical Bad Bits CIDs and early-stage indicators under investigation. The system should provide contextual information to gateway operators without silently replacing their policy decisions.

Investigate how to preserve transparent, auditable decisions while avoiding false positives and overconfident AI classifications.

Experiments

Build a searchable index of relevant content identifiers and associated policy information.
Provide contextual warnings when a gateway encounters a CID under investigation.
Distinguish known historical indicators from unverified or preliminary signals.
Test false positives, false negatives, stale information, and missing evidence.
Document how an operator can review, override, and audit decisions.

Deliverables

A working prototype, documented policy architecture, evaluation dataset, threat model, and report on accuracy, limitations, and operator control.

Stretch goal: Demonstrate integration with a self-hosted Kubo gateway and propose a reusable policy interface.

Explore

QUEST 04 · Connectivity and deployment

Intermediate–Advanced

AutoTLS, Self-Hosted Kubo, and Universal Access

Explore how self-hosted IPFS deployments can establish secure connectivity without relying entirely on publicly operated gateway infrastructure.

Challenge, deliverables & references

Investigate AutoTLS, certificate issuance for IP addresses, broker dependencies, and the requirements for secure deployment across home labs, local networks, and small enterprise environments.

Build a small test environment that evaluates certificate issuance and connectivity for self-hosted nodes. Document the operational assumptions, trust boundaries, and interoperability considerations.

Experiments

Test AutoTLS behavior in a reproducible Kubo environment.
Evaluate certificate handling for IP addresses.
Investigate broker-independent issuance as an experimental design.
Document certificate validation, renewal, and failure scenarios.
Identify opportunities for interoperability testing and specification improvements.

Deliverables

A working test environment, test cases, a deployment guide, and a report identifying implementation gaps and proposed improvements.

Stretch goal: Contribute tests or specification feedback relevant to the Kubo AutoTLS integration and broader libp2p ecosystem.

Explore

QUEST 05 · Offline-first applications

Intermediate–Advanced

Build Resilient Healthcare Applications with IPFS Meshkit

Many applications need to work when connectivity is intermittent, public gateways are unavailable, or organizations prefer to keep their infrastructure under their own control.

Challenge, deliverables & references

Use IPFS Meshkit, Kubo, Boxo, and IPFS Cluster to extend an existing healthcare application with local storage, content-addressed backups, replication, and reliable retrieval.

Choose an existing SocialCalc-based healthcare application such as Medical Invoice, Patient Sheet, Medical Notes, Patient Register, Immunization Log, Medical Requisition Form, or Headache Log.

Demonstrate how an application can save data locally, produce content-addressed artifacts, synchronize or replicate selected data, and recover from interruptions.

Experiments

Integrate Meshkit with an existing application.
Store and retrieve JSON, PDF, or CSV artifacts through IPFS.
Test local persistence during Internet outages.
Replicate selected content across Kubo nodes and coordinate pinning with IPFS Cluster.
Evaluate client-side encryption and access-control boundaries.
Test recovery, synchronization, and data integrity after reconnection.

Deliverables

A working application, documented storage architecture, offline and recovery tests, and a reproducible demonstration using local or multi-node infrastructure.

Stretch goal: Deploy a portable IPFS setup—a small “IPFS Suitcase”—that can be used in classrooms, clinics, workshops, or other low-connectivity environments.

Use synthetic or appropriately de-identified data for demonstrations; do not publish sensitive patient information to a public network.

Explore

QUEST 06 · SocialCalc and EtherCalc

Intermediate–Advanced

Reinvent Collaborative Spreadsheets with Content-Addressed Storage

SocialCalc and EtherCalc provide a practical foundation for exploring how familiar productivity applications can benefit from decentralized infrastructure.

Challenge, deliverables & references

Build an IPFS-backed spreadsheet workflow in which documents, exports, snapshots, or selected artifacts can be stored and retrieved using content identifiers. Explore how local-first editing, spreadsheet collaboration, and decentralized persistence can work together.

Use SocialCalc-based healthcare and business applications as real-world test cases. Investigate the difference between storing a spreadsheet artifact on IPFS and implementing a complete collaborative editing or synchronization protocol.

Experiments

Integrate a SocialCalc or EtherCalc workflow with IPFS Meshkit or Kubo.
Store spreadsheet exports and snapshots as content-addressed artifacts.
Retrieve and verify stored documents using their CIDs.
Test local-first editing and persistence during network interruptions.
Explore versioned snapshots, recovery, and sharing workflows.
Compare centralized synchronization with an IPFS-backed storage layer.

Deliverables

A working spreadsheet application, an IPFS integration, a documented data and synchronization model, and a demonstration showing content integrity and recovery.

Stretch goal: Prototype a portable spreadsheet workspace that can be exported, backed up, and transferred between independently operated IPFS nodes.

Explore

QUEST 07 · AI, MCP, and spreadsheet automation

Intermediate–Advanced

Build IPFS-Native SocialCalc AI Agents

Turn spreadsheet automation into a practical experiment in decentralized application architecture.

Challenge, deliverables & references

Use the SocialCalc AI framework and its MCP server to build an AI-assisted workflow that can read, transform, and generate spreadsheets while using IPFS for content-addressed storage and retrieval.

The existing SocialCalc tooling includes spreadsheet operations exposed through MCP, an AI agent architecture, and support for natural-language spreadsheet actions. The expedition challenge is to connect these capabilities with portable storage, verifiable artifacts, and resilient execution.

Experiments

Integrate the SocialCalc AI framework with IPFS Meshkit or Kubo.
Use MCP tools to read or modify spreadsheet workbooks.
Store source documents, generated reports, or snapshots using CIDs.
Record inputs, outputs, and relevant execution artifacts for reproducibility.
Test recovery after interrupted workflows.
Explore how agents can exchange spreadsheet artifacts through content identifiers rather than depending on a single storage service.

Deliverables

A working AI-assisted spreadsheet workflow, an MCP integration, a content-addressed artifact store, and a reproducible demonstration of document generation or transformation.

Stretch goal: Build an IPFS-backed invoice agent that generates spreadsheet-based invoices and exports, verifies stored artifacts, and supports portable storage across deployments.

Explore

QUEST 08 · Distributed data and AI infrastructure

Intermediate–Advanced

Build a Content-Addressed Repository for Models, Datasets, and Agent Artifacts

AI and research workloads frequently download large models, datasets, embeddings, and other artifacts repeatedly, even when multiple machines need identical content.

Challenge, deliverables & references

Build a local content-addressed repository using IPFS to publish, retrieve, verify, and replicate these artifacts across commodity machines.

Explore how nearby IPFS nodes can improve reuse and resilience when external connectivity is slow or unavailable. Investigate how content identifiers can help identify exact artifact versions and verify that retrieved data matches the expected content.

You can also extend the experiment to agent-generated files, execution receipts, and other artifacts that need to move across different storage environments.

Experiments

Publish models, datasets, or research artifacts using CIDs.
Retrieve artifacts from local or nearby nodes.
Replicate selected artifacts across machines.
Measure retrieval latency, bandwidth consumption, and storage overhead.
Test operation during Internet interruptions and node failures.
Document model provenance, versioning, and integrity-verification workflows.

Deliverables

A working artifact repository, reproducible benchmarks, a deployment guide, and a report describing the benefits and limitations of the approach.

Stretch goal: Integrate an agent workflow that retrieves verified artifacts, generates a result, and stores the output as a content-addressed artifact for subsequent retrieval.

Explore

// how the quests fit together

Modular by design

Start with a working IPFS application, progress into storage engineering and networking, then apply the resulting infrastructure to collaborative software, healthcare, and AI workflows.

Theme	Quests	Intended outcome
Foundations and libraries	01	Understand IPFS across implementations and languages
Replication and resilience	02	Build and measure fault-tolerant storage
Gateway security and policy	03	Prototype transparent, operator-controlled gateway decisions
Secure self-hosting	04	Investigate AutoTLS and deployment reliability
Offline-first applications	05	Demonstrate local storage and recovery in healthcare workflows
Collaborative applications	06	Make SocialCalc/EtherCalc artifacts portable and verifiable
AI and spreadsheet automation	07	Connect MCP, AI agents, and content-addressed artifacts
Distributed AI infrastructure	08	Replicate and verify models, datasets, and execution artifacts

// the implementation landscape

One ecosystem, multiple languages

The expedition encourages contributions across the IPFS and libp2p ecosystem. Participants should select the implementation that best fits their experiment rather than assume that every feature exists in every language.

Go — Kubo and Boxo

Explore Kubo as an IPFS node and Boxo as a collection of reusable IPFS components. Focus on storage, UnixFS, retrieval, integration, replication, and testing.

JavaScript — Helia and Meshkit

Build application-level IPFS integrations, browser or Node.js workflows, offline-first interfaces, and integrations with SocialCalc, MCP, and AI agents.

Python — libp2p and application integration

Investigate py-libp2p for peer-to-peer networking experiments and build Python-based tooling, benchmarks, or application integrations. Where a Python IPFS implementation does not expose a required capability, document the gap or interoperate with an existing IPFS node through a defined interface.

Multi-node infrastructure — IPFS Cluster

Use IPFS Cluster where appropriate to coordinate pinning and replication across multiple IPFS nodes. Test failure recovery, retrieval, and operational trade-offs.

Shared testing — interoperability and reproducibility

Develop reusable tests and measurements that can help implementations and application developers understand connectivity, retrieval, performance, and failure behavior.

// recommended participation format

Learn. Build. Test. Contribute.

1. Learn

Understand the application, library, protocol, and problem you are addressing. Read the relevant code and identify the behavior you want to improve.

2. Build

Create a minimal working prototype, integrate an existing application, or reproduce a known limitation.

3. Test

Use reproducible experiments, multi-node deployments, interoperability tests, or benchmarks to establish what works and where it fails.

4. Contribute

Publish your results, document trade-offs, submit upstream pull requests, and share follow-up ideas with maintainers and fellow builders.

// what does completing a quest mean?

One meaningful, verifiable contribution

✅ Original contribution

Solve a meaningful technical problem, build a novel application, improve an existing implementation, or establish a new research result. Simply following a tutorial is not sufficient.

✅ Open-source repository

Provide working code, clear setup instructions, documentation, tests, and links to relevant upstream repositories or pull requests.

✅ Research or technical report

Explain the problem, architecture, implementation, experiments, results, limitations, and next steps.

✅ Working demonstration

Show your application, replicated storage network, offline-first workflow, gateway prototype, AI integration, or interoperability test in action.

✅ Evidence of resilience

Where relevant, measure retrieval latency, replication overhead, availability, recovery after node failure, behavior during network partitions, or operation under constrained connectivity.

✅ Contribution to the ecosystem

Document what others can reuse. Submit upstream fixes where appropriate, share findings with maintainers, and identify opportunities for follow-up work.

Participants can work individually or in teams. Each submission should identify contributors and their respective roles.

For healthcare demonstrations, use synthetic or appropriately de-identified data. Never publish sensitive patient information to a public network.

// evaluation criteria

Research depth meets practical impact

Submissions will be evaluated on:

Originality and clarity of the problem.
Technical quality and architectural soundness.
Quality of implementation and testing.
Reproducibility of experiments and results.
Resilience, interoperability, and practical impact.
Documentation and quality of the demonstration.
Potential for upstream adoption or broader ecosystem use.

A strong submission connects a clearly defined problem to a credible technical solution supported by working code and evidence.

Not every quest needs to end with merged code. A well-supported research report, useful interoperability test, or reproducible prototype can be equally valuable.

// pick a quest

Ready to build resilient storage?

Choose a quest, introduce yourself to the community, and collaborate with developers and maintainers during the IPFS Planet Expedition.

Bring an existing application, reproduce a storage problem, experiment with Kubo or Helia, improve a library, or explore how SocialCalc and AI agents can use IPFS in practice.

You do not need to be an IPFS expert to participate. Start with a small experiment and grow it into a contribution that others can reproduce and build upon.

Build. Replicate. Verify. Recover. Contribute.

Join the Resilient Storage @ Devcon’s Bunkers track and help make decentralized storage useful in the real world.

Join the Resilient Networking Hub.
Explore the IPFS Planet quests.
Contribute to IPFS and libp2p.
Bring a SocialCalc, healthcare, or AI application to the expedition.

Join the Expedition Resilient Networking Hub

// beyond mumbai

Stay connected. Build together.

Continue your work with the IPFS, libp2p, and decentralized storage communities before, during, and after the expedition.

Explore the ecosystem

IPFS Planet Expedition · Resilient Storage @ Devcon’s Bunkers

A nine-day expedition into decentralized storage, resilient applications, and open infrastructure.

Devcon 8 Mumbai · October 31 – November 8, 2026

Build the storage layer of the decentralized internet.