Immutable Document Link
Project Information
| Field | Value |
|---|---|
| Project Name | Sofia |
| Co-founders | Chauche Samuel, Saint-Joannis Maxime |
| Product | Download |
| Community | X / Discord |
| Atom & Claim | Sofia Atom / Sofia Claim |
| Links | Colonny - Github - Product Roadmap - Website - Demo - Gnosis Safe |
1. Applicant & Project Overview
Summary
Sofia is a chrome extension that empowers people and organizations to capture, verify, share and amplify knowledge across the web. It transforms how you interact with web content, allowing you to truly own your digital experience.
This way, you can have recommendations built from the information you choose to share and spread some of your knowledge indirectly with your trusted ones.
Project Category
- Agent / AI Context / Consumer App
- Knowledge Graph / MCP
Elevator Pitch
We all know the feeling of the âAlgorithmic Trap.â You spend hours online, yet youâre fed content that serves the platform, not you. You rely on Amazon reviews from strangers instead of recommendations from friends.
Sofia is here to fix the broken relationship between users and their data.
Sofia is a tool that facilitates the transition of reclaiming your digital footprint. By analyzing your browsing history, Sofia builds a secure, Personal Knowledge Graph. We then use a dedicated AI to unlock the true power of that data.
Imagine an internet where your feed isnât decided by a black-box algorithm, but by your own interests and the trusted signals of your network. Sofia turns your passive browsing into active, verifiable knowledge. Itâs about giving everyone the power to stop renting their attention and start owning their digital identity.
Origin Story
We met at The Hacking Project in February 2025 with a shared goal: becoming web developers. Both passionate about electronic music, we immediately hit it off.
However, the real spark for Sofia didnât come from a line of code, but from a moment of profound frustration. In June, during a brainstorming session for our final project, Sam hit a breaking point. He realized he had just lost 30 minutes doom-scrollingâconsuming ads, Reels, and random postsâwithout learning anything new about his friends or the world.
This sparked a critical realization: Why doesnât the internet feel like âhomeâ?
We asked ourselves why we are constantly fed content designed to keep us glued rather than enriched. Why do we rely on reviews from strangers on Amazon or Netflix, when we canât easily access recommendations from the people who know us best?
A global realization is occurring: the current web model is broken and unsustainable. We decided to change the game. We didnât just want to build another app; we wanted to fix this issue and onboard millions of people into a new paradigm. Sofia was born from this desire: to stop being passive consumers of algorithms and start using our own browsing data to build a personalized, decentralized, and authentic knowledge graph.
Notable Traction or Achievements
- Mainnet Deployment: Successful migration and deployment of Sofia on the Intuition protocol mainnet.
- Verifiable Computing (TEE): Deployment of the first Trusted Execution Environment (TEE) machine on the Phala Network, marking a key step toward decentralized and secure computing.
- Stable Environment: Implementation of a fully managed environment, ensuring the stability and scalability required for the Alpha phase.
- Sofia Fee Proxy Contract (link) â A custom smart contract deployed on Intuition Mainnet that collects a 0.1 TRUST + 5% fee on deposits, seamlessly integrated with the MultiVault protocol, the contract architecture has been open-sourced as a reusable template for developers to deploy their own fee proxy on Intuition (Github).
- AI Recommendation Engine: Integration of Gaianet to power a recommendation engine capable of analyzing usersâ on-chain activity and providing intelligent, contextual suggestions.
- 5 Specialized Agents: Development and consolidation of the agent stack, including:
- Bookmark (theme extraction agent for importing bookmarks and translating them into signals (triples) in one click)
- Tab Analysis
- Real time navigation analysis
- Chatbot
- Suggestion (agent analysis)
- Trust this site component: Introduction of a major feature allowing users to instantly verify a web page and view the underlying trust score, marketcap, triples and atoms related.
- Branding: Invested time in branding, no major update planned for the next quarter.
- UX: Integrated into the core agile development process, ensuring a user-centric approach from conception to launch. Focused on delivering an intuitive and engaging user experience that drives key business outcomes
- Trust Circle: Designed as a tier above the standard Follow, the Trust Circle reinforces mutual confidence and activates our new AI-driven protocol for seamless information sharing between trusted users.
- Blog: We established Sofia Chronicles, a dedicated blog to maintain transparency and community engagement. To date, we have published 10 weekly changelog covering development updates, tutorials, and project insights.
Process and Community
- Public Presentation: Sofiaâs first public demo at an Intuition X Space event in front of 8K viewers and 9 submissions to the early access program.
- Community: Creation of the Sofia Discord server (41 users) with MEE6 licence. 83 followers on X.
- AlphaTesting: Private Onboarding of the first 5 alpha testers, paving the way for the Beta phase.
Project Management
- Internal Governance: Adoption of Agile methodology and Core Protocols to improve communication, decision-making, and efficiency within the development team.
Current Users or Early Testers
- Zet
- Jade
- Max (Wieedze)
- Sam (Passive-Record.box)
- W00ds
- Caleb
2. What we are building
Problem Statement
The modern internet is exhausting: algorithms are optimized for user retention and generate advertising rather than relevance, creating an environment where personalized content is untrustworthy and information lacks verification. Users lack a digital space that truly reflects their identity.
Proposed Solution
Sofia reverses the attention economy by prioritizing Trust and Relevance. Sofia uses on-chain data and local AI (Gaianet) to provide context-aware recommendations, enabling users to verify information and collaboratively help to build the decentralized knowledge graph which can be used by agents and for end-users.
Stage of Development
Alpha Launch is initiated, slowly ramping up (The core features are deployed on the Intuition mainnet, secured via Phala Network, and currently undergoing intense testing with real users in preparation for a wider public release.)
Technical Architecture Overview
Integrations or Dependencies
- Gaianet: Nodes, LLM
- Phala: TEE
- ElizaOs: Agent Framework
- Intuition Indexer: GraphQl API
- Intuition Protocol
- Plasmo: Extension Framework
- Frontend libraries
Privacy Considerations
Sofia is fundamentally designed to enhance digital security through decentralization and verifiable computation. This is achieved by deploying its first TEE machine on the Phala Network, and powering the agent framework with a node link to a model ensuring that computations are both secure and verifiable, minimizing reliance on centralized servers.
Furthermore, the core Trust system allows users to verify web pages and link personal insights to verifiable on-chain data, promoting data integrity and a community-driven layer of security. Secure OAuth integrations (e.g., with YouTube, Spotify) were implemented with a dedicated, clean structure to protect user credentials.
3. Team & Execution Ability
Team Backgrounds
Samuel Chauche:
- 6+ years as a Support Engineer through several companies - Credit Suisse, Airbus, Michelin, Salesforce
- THP Graduate - intensive fullstack development program
- Intuition Contributor - documentation, UI, ERC-8004
- 3rd place in the Intuition hackathon - prediction market based on the number of transactions passing through the intuition testnet.
- 4 months full-time on Sofia then 50-50 between Sofia and Intuition (documentation, UI, ERC-8004)
- Part of Rings of Neptunes agency as electronic music producer (ambient/experimental). Link
Maxime Saint-Joannis:
- THP Graduate - intensive fullstack development program
- 4th place in the Intuition Hackathon
- Full-time on Sofia since June
- Professional Music Producer since 10 years (Radikal Moodz) Video Link
Team Structure and Roles
| Name | Role |
|---|---|
| Samuel Chauche | Co-founder |
| UI/UX Developer | |
| Marketing/Communication | |
| Maxime Saint-Joannis | Co-founder |
| CTO & AI Infrastructure Architect |
Advisors or Collaborators
- Zet
- Saulo
- W00ds
- Caleb
- Shifou
- Jade
Expected Future Contributors
- CM, CT - Managing X communications, content creation and Discord development - Full time
- Back-end Developer - Part time
4. Grant Request & Milestones
While our initial plan for this grant request detailed only the next three months, the enthusiasm and strategic necessity of our growing project demanded we develop a comprehensive one-year vision. This extended roadmap is fundamental to our overarching goal: bootstrapping a truly decentralized organization (DAO) driven by a large, active community of builders.
This commitment to transparency is central to our ethos of âbuilding in public,â which we uphold through platforms like the Sofia Chronicles blog and is validated by the early traction in Sofia atoms and Sofia Claim on the portal. We believe this open communication is crucial for converting early enthusiasm into a robust, dedicated community that will eventually govern the protocol.
Beyond securing the necessary initial support, the fundamental goal of this document is to initiate a deeper, more synergistic partnership with Intuition Systems. This collaboration is key to accelerating our transition toward a decentralized governance model. We are prepared to fully open Sofiaâs roadmap to the Intuition team, and in turn, seek a reciprocal transparency regarding Intuitionâs priorities, thus establishing an open, two-way collaboration that ensures maximal long-term strategic alignment for both ecosystems and, crucially, for the future DAO.
Below, we detail our one-year plan, structured into three strategic stages. For a look beyond this period, please refer to the dedicated chapter on âSustainability & Long-Term Visionâ
Consistent with our commitment to strategic alignment, we defer to the Intuition team to adjust the scope and funding amount of this first milestone, ensuring it corresponds to the highest strategic priorities.
The final allocation split between USD and $TRUST token remains flexible at this stage and is a topic we look forward to discussing directly with the Intuition team.
Monthly Budget Breakdown (USD)
| Item | Provider | Amount |
|---|---|---|
| Nodes Gaianet LLM Qwen 2.5 72B | ||
| GPU: NVIDIA RTXâą 6000VRAM: 48GBRAM: 128 GB2 x 1.92 TB NVMe | Hetzner | $940 |
| Framework ElizaOS | ||
| XLarge TDX Instance 8 vCPU - 16GB RAM (128GB) | Phala | $410 |
| Core Dev Team | ||
| Covers 2 FTE (Full Time Equivalent) specialized engineers including taxes | Maxime Saint-Joannis / Samuel Chauche | $7,000 |
| DAO Operation | ||
| (event, CM, CT, Content creation, discord management, X account management) | Community | $5,000 |
| Monthly total | $13,350 |
Milestone 1: Alpha Launch
Estimated Duration: 3 months
FTE: 2 FTE (Full Time Equivalent)
Costs: $40,050 USD
Deliverables
| Number | Deliverable | Specification |
|---|---|---|
| 1.1 | Intuition MCP Connection | Full implementation of the Chatbot connected to the Intuition MCP Server via API with direct possibility of action. |
| 1.2 | Public Alpha Release | Deployment of the Chrome extension (Alpha) / Chrome Web Store. |
| 1.3 | Community Setup | Initial deployment of the Community Management |
| 1.4 | Documentation | Detailed documentation covering core protocol architecture and installation guide for developers. |
| 1.5 | Feature | - Untrust Website component- Publish BookMark as list- Share List with Trusted User- Quest and reward system- Signal Curve display |
| 1.6 | Release | Publication in Chrome Extension Webstore (Private release) |
| 1.7 | Governance Blueprint | Finalized the DAO creation on Colonny, including proposal structure and voting parameters. |
| 1.8 | Metrics | Develop a dashboard linked to the fee proxy to display smart contract metrics. |
Success Criteria
M1: Core Infrastructure & Alpha Launch
- Colony and Multisig operational
- 10+ community member received 10k+ USDC
- 100 weekly users
- +5000 transactions through the proxy
- User satisfaction rate 7/10
Stage 2 â Beta Launch & Scaling
| Number | Deliverable | Specification |
|---|---|---|
| 2.1 | R&D | Implementation if possible of Account Abstraction (ERC-4337) within the extension for improved user experience and subsidized gas fees. |
| 2.2 | TEE | Agent Scale for better task management |
| 2.3 | TEE | Split agent in different TEE machine |
| 2.4 | GPU | Augment the LLM size for better computing |
| 2.5 | GPU | Create a new node for simple task to lighten the main |
| 2.6 | Feature | - Start Linked Triples economic- 3 New importation modules (e.g: Google Maps, Opentable, Booking)- Drag and Drop external files (pdf, picturesâŠ) and translate information into triples- Show me similar website |
| 2.7 | Public Beta Release | Launch of the Public Beta version |
| 2.9 | Growth | Establish a distinctive and recognizable lore, visual design, and brand identity that are validated through user feedback and engagement metrics. (Marketing Material) |
| 3.1 | Founding | Grants submission for Gaianet |
| 3.2 | Founding | Grants submission for Elizaos |
Success Criteria
Stage 2: Beta Launch & Scaling
- 50+ contributor on DAO
- 10,000+ X subscribers
- 2+ Grants approved
- 1,000 weekly users
- +100,000 transactions through the proxy
- +10,000 transactions from the import modules
- 50% Reduction of latency for complex queries after LLM augmentation
- User satisfaction rate 8/10
Stage 3 â Economic & Governance
| Number | Deliverable | Specification |
|---|---|---|
| 3.1 | Linked Triples Economics | Implementation and testing of the Linked Triples economic |
| 3.2 | GPU | Gaianet Domain ecosystem set up |
| 3.3 | GPU | Divide tasks in differents nodes for better load balancing |
| 3.4 | Feature | - Notification system (Pushchain)- Sofiaâs trust score on Google results- Reward system for linked triple economics usage |
| 3.5 | Governance | Incorporating community ideas into the roadmap |
| 3.6 | Tokenomics | Sofiaâs tokenomics design and launch |
Success Criteria
M3: Economic & Governance
- Target of 10,000 weekly users
- 1M+ transaction
- Users actively participate in the Linked Triples economic mechanic
- Sofiaâs TGE
Stretch Goals
Linked Triples Economic Model
Develop a treasury development model leveraging Intuitionâs knowledge graph logic, nested triples, and bonding curve mechanisms. This model will enable Sofia to generate recurring treasury accumulation through the creation and curation of valuable knowledge graph relationships while rewarding users who actively contribute to and utilize the Linked Triples via an equitable reward mechanism. The overall design will maintain alignment with user interests and platform growth.
Sofia Fee Structure: Sofia charges a small fee through a proxy contract on all blockchain operations to support the development and maintenance of the platform. For creation operations (atoms and triples), we charge a fixed fee of 0.1 TRUST for deposits and we charge 5% of the deposit amount. All collected fees are sent directly to a Gnosis Safe multisig wallet managed by the Sofia core team. These fees help us continue improving the Sofia experience while keeping the platform sustainable. As Sofia grows and reaches wider adoption, we plan to progressively reduce these fees over time.
Linked Triples Architecture Suggestion
The idea here is that Sofia always participates in bonding curves.
- Step 1: Sofia creates triples from navigation analysis (Youtube) (e.g. I like Donald Trump)
- Step 2: The user invests in the triplesâSofia then create a linked triples (e.g. Youtube knows Donald Trump)
- Step 3: Another user invests in the same linked triples > Sofia withdraws a part her initial position > Then places another position (linear curve)
Web 2 User Attraction
- Wrap all the interaction with the Intuition block-chain in an abstraction layer so the user doesnât have any feeling of interacting with a complex technology (e.g: delegation through ERC-4337).
- Import all the data from the userâs favorite platform (X, Google Maps, OpenTable, Booking, AirbnbâŠ).
- Import external data from personal life (books, doc, pictureâŠ).
Community-Driven Governance
Implement a transparent decision-making DAO where the Sofia community actively participates in product roadmap decisions through structured feedback and decentralized voting.
We want to build a meritocratic DAO as the core decision-making and governance interface of our platform. Instead of relying on purely financial influence, our governance system rewards meaningful contributions such as participation, curation, knowledge creation, and trust-building as the real drivers of influence.
Deep Intuition Ecosystem Integration
Become a key contributor to the Intuition ecosystem by:
- Building native integrations with Intuitionâs core infrastructure
- Contributing to protocol development by driving adoption to Intuition
- Establish partnerships with complementary Web3 projects
Key Dependencies
Intuition Protocol Stability: Sofiaâs core functionality relies entirely on the stability, performance, and feature set of the underlying Intuition mainnet and indexer. To ensure maximum reliability and contribute directly to the protocol, we plan to become active validators by hosting one or several Intuition nodes.
Phala Network Performance: The successful deployment and scalability of the TEE machines depend on the sustained performance and availability of the Phala Network.
Community Adoption: The core value proposition is directly dependent on active user participation to create and validate the verifiable knowledge graph.
Gaianet: Decentralized infrastructure of open source model through domain rental and node hosting. This ensures secure and anonymous model usage for all users.
Risk Assessment
Technical vulnerability: risk of bugs among partners (e.g major upgrade of our dependencies, discontinuation of one or more services)
Mitigation: Establish a genuine partnership, thorough security audits before launch, gradual deployment (Alpha â Beta) (Blue â Green)
Adoption and engagement: risk that users will not use Sofia, or will use it very little, because the technology is still not widely understood.
Mitigation: Educational focused on the unique advantage in terms of relevance. Creation of videos for presentation/training purposes
How This Work Compounds Value for the Network
- Sofia aimed to become the flagship application for the Intuition protocol, allowing users to intuitively create a claim to improve the knowledge graph based on his habits.
- By creating claims and identity from web navigation habits and personal human input we help create a real data network valued by humans and helpful for agents.
- This web-navigation analysis combined with credibility signals allows the user to build a certification map that helps future users to navigate with greater awareness.
- By encouraging the creation of claims, which helps to strengthen the integration of new data.
- Sofia will allow users to intuitively interact with the knowledge graph of their trusted circles, while still preserving the full value of the broader, globally accessible knowledge graph.
- Finally, by highlighting the secure and verifiable nature of personalized AI through AgentRank (ERC-8004), Sofia and Intuition prove the usefulness of a decentralized and verifiable knowledge graph that will revolutionize the way information is exchanged between users.
5. Intuition Ecosystem Alignment
Why Intuition?
We chose Intuition for transparency, ethics, innovation, and open-mindedness. Distributed verification and storage guarantee the portability and inviolability of the knowledge graph. Intuition is creating a new way to certificate the web and how trusted information can come in your private space.
Which Intuition Primitives You Will Use
- Atoms,
- Triples
- Nested Triples
- Lists
- MCP
- Knowledge Graph
- AI context
- Permission graph
Why This Must Be Built on Intuition
This project can only be built on the Intuition Protocol, because Sofiaâs mission is fundamentally inseparable from Intuitionâs core principles and its ability to enable a decentralized, verifiable web.
Verifiable Knowledge Graph: Intuition provides the essential infrastructure for handling triples and atoms, the semantic building blocks required to create Sofiaâs community-validated knowledge graph. This is necessary for moving beyond proprietary, siloed and unverified data structures.
Decentralized Trust and Identity: The protocol provides the mainnet and indexer infrastructure needed to store and verify on-chain identity and trust metrics. Without Intuition, Sofiaâs central featureâthe decentralized âTrust this siteâ system and its connection to user activityâwould lack its foundational layer of verifiable truth.
Data Integrity: Intuition enables the publishing of user insights directly on-chain, strengthening the link between user behavior and verifiable blockchain data. This prevents data manipulation and ensures the integrity required to challenge centralized algorithms.
Recommendation Layer: Sofiaâs recommendations are built directly from the knowledge the user has intentionally written to their personal graph. Working as a personal suggestion algorithm.
Links
- Github
- Website
- Sofia Chronicles
- DAO Sandbox
- Alpha tester Guide
- Visual Identity Research
- Figma Visual Identity
- Initial Figma project
- Whimsical (Initial Brainstorm)
How Do You Plan to Integrate TRUST, the Native Currency of Intuition Network
Trust is the native currency used in our application to interact with the Intuition protocol. The user can influence the recommendation agent by weighting the claims in their profile, which adjusts the level of interest assigned to each topic.
Users can follow others and create trust links by issuing claims such as âI trust (userX)â. This feature allows them to track othersâ moves on various bets and intuitively join the actions taken by their trusted circle.
Finally, TRUST will be the only currency used within our application.
Data Structure Plan (Atoms, Triples, Signal, Schemas)
Predicate Upgrade Rules (Automatic Suggestions)
| To Predicate | Visits | Avg Duration | Attention | Days Active |
|---|---|---|---|---|
| master | >= 100 | > 8 min | > 0.95 | >= 30 |
| value | >= 50 | > 6 min | 0.85 | >= 14 |
| likes | >= 20 | > 3 min | 0.70 | >= 7 |
| are interested by | >= 8 | > 2 min | 0.50 | - |
| has visited | >= 1 | - | - | - |
Domain Intention Score Calculation
- frequencyScore = log(visitCount + 1) * 0.8
- durationScore = min(avgDurationMinutes / 5, 2) * 1.2
- attentionScore = maxAttentionScore * 1.5
- predicateScore = sum(predicateCount * weight) * 0.5
- recencyBonus = max(0, (7 - daysSinceLastVisit) / 7) * 0.8
- consistencyBonus = (if visits > 3 && days > 1) min(days/30, 1) * 0.5
Sofia Agent Outputs Schema
Supported Predicates
| Agent | Predicates |
|---|---|
| Sofia | trusts, master, value, like, are interested by, have visited |
| ThemeExtractor | are, like, use, watch, listen to, code in, work with, buy, play, are interested in, are learning about |
| PulseAgent | research, compare, evaluate, plan, learn, explore, consider, analyze |
Data Flow
New Schemas, Patterns, or Agent Types You Might Introduce
Nested Triples Schema for Predicate Generation
Transparent proof of predicate generation system in sofiaâs agents by creating all the relation in the protocol
Linked Triples Agents (Autonomous Graph Builder)
Purpose: Autonomously creates linked triples to build semantic chains connecting user interests to deeper knowledge relationships. When a user creates a triple, they will create three relationships or at least participate in three triples. This multiplies user interaction compared to normal activity.
Operates: Independently via MCP (Model Context Protocol)
Acts as: Sofiaâs autonomous knowledge graph expansion agent
How This Increases Network Activity or Knowledge Density
The introduction of the LinkedTripleAgent transforms Sofia from a simple triple recorder into an autonomous knowledge graph builder. Currently, user browsing generates isolated triples (User â predicate â Object) with no inter-connections, resulting in a sparse, flat graph. The LinkedTripleAgent autonomously expands each user triple by querying related concepts via MCP and creating 1-3 linked triples with semantic relationships (is related to, part of, leads to, knowsâŠ), plus reverse links back to the userâmultiplying triple creation and blockchain operations.
I confirm that all information submitted is accurate and that I intend to deliver the listed milestones