Blockchain Research Platforms

TankTerminal and TankWork represent innovative approaches to AI-blockchain integration, offering both no-code and developer-focused solutions for building AI agents that can interact with computers. TankTerminal stands out by providing a user-friendly platform where anyone can deploy AI agents with voice control capabilities, making it accessible for non-technical users to leverage AI automation on their desktop systems.The framework excels in combining AI capabilities with computer automation, allowing for sophisticated interactions between blockchain applications and desktop environments. TankWork's open-source nature provides developers with the tools to create custom AI agents using Python and PyQt, enabling deep integration possibilities between AI systems and blockchain applications. This makes it particularly valuable for developers looking to build automated trading systems, blockchain monitoring tools, or AI-powered DApp interfaces that require desktop-level access and control.

Spore Platform stands out as an innovative AI integration solution on Solana, leveraging blockchain technology to create an autonomous ecosystem for AI evolution. The platform utilizes Trusted Execution Environment (TEE) verification to ensure secure and validated AI agent interactions, while implementing token incentives to drive the evolution and reproduction of AI entities in a decentralized manner.The platform represents a groundbreaking approach to AI development on Solana, where AI agents can autonomously evolve, learn, and reproduce based on predefined parameters and success metrics. Through its token-based incentive system, Spore creates a self-sustaining ecosystem where successful AI behaviors are rewarded and propagated, while the TEE verification ensures the integrity and security of all AI operations. This combination of blockchain technology and artificial intelligence opens up new possibilities for decentralized AI development and evolution.

The Pippin Framework represents a sophisticated integration of AI capabilities with blockchain technology on Solana, offering advanced autonomous agent development tools. By implementing BabyAGI's iterative task planning alongside enhanced memory systems, Pippin enables developers to create AI agents that can interact intelligently with blockchain data and smart contracts. The framework's modular architecture allows for seamless integration of AI-powered decision making and automated task execution within the Solana ecosystem.The framework's character definition protocols and dynamic skill management system make it particularly valuable for developing AI-enhanced dApps on Solana. Through its persistent memory evolution and sophisticated logging mechanisms, Pippin enables AI agents to learn from their interactions with the blockchain, adapt their strategies in real-time, and maintain consistent performance across various blockchain operations. This combination of AI capability expansion and blockchain integration makes it a powerful tool for developers looking to create intelligent, autonomous applications on Solana.

Centience Worlds represents an innovative fusion of AI and blockchain technology on Solana, offering an advanced media workstation for exploring and generating AI-powered content. The platform leverages language models to create interactive multimedia experiences, allowing users to seamlessly integrate text, image, video, and audio synthesis in a unified creative environment.The platform stands out in the AI-blockchain integration space by providing a sophisticated workstation that enables real-time exploration of language model capabilities. Users can experiment with different AI models, fine-tune parameters, and generate cross-modal content while maintaining the security and transparency benefits of blockchain technology. The interactive nature of Centience Worlds makes it particularly valuable for creators and developers looking to push the boundaries of AI-generated media on Solana.

Whispers Protocol is revolutionizing the integration of AI with blockchain technology on Solana by creating a sophisticated agent orchestration system. The protocol intelligently matches and coordinates AI agents for various crypto operations, leveraging Solana's high-speed infrastructure to enable real-time agent selection and task execution. This dynamic system allows for seamless integration of multiple AI agents working in concert to handle complex DeFi operations, trading strategies, and other blockchain activities.The platform features an innovative marketplace where developers can deploy their specialized AI agents, creating a rich ecosystem of purpose-built tools for different blockchain operations. Each agent can be evaluated based on historical performance, specialization, and efficiency metrics, ensuring optimal selection for specific tasks. The protocol's ability to coordinate multiple agents simultaneously while maintaining security and performance makes it a groundbreaking solution for AI-blockchain integration on Solana.

ViralMind Training Gym represents an innovative fusion of AI and blockchain technology on Solana, creating a unique platform where users actively contribute to training artificial intelligence models through gamified computer tasks. By completing specific actions and generating valuable training data for Large Action Models (LAMs), users are rewarded with $VIRAL tokens, creating a symbiotic relationship between human participants and AI development.The platform leverages Solana's high-speed infrastructure to process and validate user contributions in real-time, while smart contracts ensure transparent and immediate reward distribution. This AI-blockchain integration demonstrates a novel approach to crowdsourcing AI training data, where each user interaction is recorded on-chain and contributes to the development of more sophisticated AI models that can better understand and replicate human actions.

Franklin X represents a pioneering advancement in AI-blockchain integration on Solana, offering autonomous agent capabilities through the Virtuals platform. As the first AI agent in the ecosystem, it enables sophisticated automation of blockchain interactions using natural language processing and intelligent decision-making systems. The platform can interpret user instructions and convert them into executable blockchain transactions while maintaining security and validation checks.Franklin X's AI capabilities extend to monitoring market conditions, evaluating smart contract states, and executing complex transaction sequences automatically. The system's advanced language processing allows users to interact with blockchain functionality through conversational interfaces, making complex DeFi operations and trading strategies more accessible to users without technical expertise. This combination of AI and blockchain technology opens up new possibilities for automated portfolio management, trading strategies, and cross-protocol interactions within the Solana ecosystem.

SendAI stands at the forefront of AI integration on Solana, offering developers powerful tools to create autonomous AI agents through its Solana Agent Kit. With over 95,000 NPM downloads, the kit enables the creation of sophisticated AI agents capable of executing more than 30 different blockchain actions, from automated trading to complex DeFi operations. The platform's support for both TypeScript and Python implementations makes it accessible to a broad range of developers while maintaining high performance and security standards.The project's comprehensive ecosystem approach includes partnerships with over 40 major protocols, allowing AI agents to interact seamlessly with DEXes, NFT marketplaces, and cross-chain bridges. SendAI has demonstrated its commitment to advancing AI integration through major hackathon events offering $275,000+ in prizes, fostering innovation and development of new AI use cases on Solana. Their extensive documentation, development tools, and security features make it a standout platform for building AI-powered blockchain applications.

Bonsol stands out as an innovative ZK co-processor for Solana, providing developers with powerful zero-knowledge proof capabilities to enhance privacy and scalability in their applications. By enabling off-chain computation with on-chain verification, Bonsol allows developers to process complex calculations privately while maintaining the security guarantees of the Solana blockchain.The platform's sophisticated ZK-proof system enables developers to build applications that can handle sensitive data and complex computations without exposing the underlying information on-chain. This makes Bonsol particularly valuable for applications requiring privacy-preserving features such as private voting systems, confidential financial transactions, or anonymous authentication mechanisms. The co-processor's architecture is specifically optimized for Solana's high-performance environment, ensuring minimal latency and maximum throughput for ZK-proof generation and verification.

Brian App represents a significant advancement in AI-blockchain integration on Solana, offering a natural language interface that transforms plain English instructions into executable blockchain transactions. The AI agent demonstrates sophisticated understanding of DeFi protocols, token economics, and cross-chain operations, allowing users to perform complex operations like token swaps, bridging, and staking through simple conversational prompts rather than navigating multiple dApp interfaces.The platform's API and SDK components provide developers with powerful tools for integrating AI-powered blockchain interactions into their applications. By processing natural language inputs and converting them into structured transaction data, Brian enables unprecedented accessibility to blockchain functionality while maintaining the security and precision required for financial operations. This infrastructure layer helps bridge the gap between traditional user experiences and blockchain complexity through sophisticated language processing and transaction generation capabilities.

Chainbase's Theia Crypto World Model represents a cutting-edge integration of AI and blockchain technology on Solana, enabling sophisticated data analysis and predictive modeling. The AI model is specifically designed to understand and process blockchain data, supporting applications across gaming, DeFi, security, and asset management through advanced simulation and reasoning capabilities.The platform leverages machine learning to analyze patterns in blockchain data, providing valuable insights for decision-making and risk assessment. Theia's ability to process and learn from native blockchain environments makes it particularly powerful for developing AI-driven applications, from automated trading strategies to security threat detection. The model's comprehensive understanding of crypto ecosystems enables it to support complex use cases like market prediction, anomaly detection, and portfolio optimization.

SwarmZero is an AI agent platform positioned within the Solana AI ecosystem, with tooling designed for blockchain use cases. The SDK enables developers to create agents that perform blockchain data analysis, DeFi metrics monitoring, smart contract and wallet risk screening, and multi-chain crypto trading automation, including Solana-specific workflows. The platform supports nine LLM providers, letting developers route agents to different models based on cost or capability. SwarmZero participated in Outlier Ventures' AI x Crypto Base Camp accelerator, a program focused on decentralized AI infrastructure, placing it at the intersection of AI agent development and blockchain technology.

Daydreams is infrastructure for autonomous AI agents to participate in on-chain commerce without human intermediation. Its Task Market connects requesters and fulfilling agents through USDC-settled escrow contracts, supporting five competition modes — Bounty, Claim, Pitch, Benchmark, and Auction — to match incentive structures to different task types. By mid-2026, the Task Market had attracted over 850 competing agents on Base alone, with a single $100 task drawing 119 competing agents and the platform processing 1.4 million transactions in a two-week period. The underlying payment standard is x402, an open protocol that embeds USDC micropayments into the HTTP request/response cycle. Agents discover funded work, compete for tasks, complete them, and collect USDC without a human operator approving each step. The Dreams Router extends this settlement layer across Base, Solana, and Starknet, and doubles as an AI model gateway providing OpenAI-compatible access to models from multiple providers, with inference costs settled per token via the same x402 micropayment rails.

FactCheck.fun places AI agents at the core of its fact-resolution mechanism, using them to autonomously browse the web, locate supporting or contradicting evidence, and assign confidence scores on a 0–100 percent scale. On Central Limit Order Book markets, trading halts automatically when confidence reaches 90 percent and the market enters a 24-hour resolution window settled via Volume-Weighted Average Price with a randomly selected final block to prevent manipulation. On Dynamic Pari-Mutuel markets, a panel of AI agents reviews submitted evidence and renders a final verdict. Founded by Prithvir Jhaveri, a former AI product manager backed by Alliance DAO, the platform positions AI-assisted verification as a credibly neutral alternative to editorial fact-checking. By combining machine evidence-retrieval with crowd-sourced financial stakes, FactCheck.fun generates a real-time, publicly observable confidence signal that traditional editorial processes cannot produce, making truth-seeking a financially incentivized and algorithmically grounded activity.

Auto.fun implements AI-blockchain integration by making the autonomous agent an integral part of the token launch rather than a feature to be built afterward. The no-code builder translates plain-language instructions into elizaOS agent configurations, allowing builders without programming backgrounds to deploy agents that post on X, respond to holders in Discord, monitor wallet activity, and execute DeFi strategies. Agents operate from the moment of launch without creators needing to maintain a server or write integration code. The platform runs on elizaOS, the open-source AI agent protocol built by Eliza Labs, which provides the execution layer for all autonomous operations. ElizaOS v2.0.x introduced an OWNER+AGENT identity architecture separating on-chain identity from operator credentials, a design that flows through to every auto.fun deployment. The infrastructure has proven generalized beyond auto.fun itself: a second launchpad, waifu.fun, was subsequently built on elizaOS v3 agents, demonstrating the composability of the underlying AI-blockchain stack.

Zerion released the Zerion CLI in May 2026, an open-source toolkit that allows AI agents — including Claude, Cursor, and OpenAI Codex — to query portfolio data and execute blockchain transactions via natural language commands. The CLI is built on the Zerion API, supports 40-plus EVM chains and Solana, and covers portfolio tracking across 8,000-plus DeFi protocols, token swaps, bridge routing, transaction signing, and profit-and-loss calculations. An MCP server ships with the CLI, enabling direct integration with Claude, Cursor, Windsurf, and VS Code. The CLI is extensible through a modules system called Skills, allowing developers to add new protocol integrations and workflows. At launch, MoonPay, Uniswap, Polymarket, Monad, and LI.FI contributed Skills to the catalog. This positions Zerion's data infrastructure as a programmable layer for autonomous financial agents managing portfolios and executing strategies across multiple chains without human intervention at each step.

Natix Network integrates on-device AI with Solana blockchain incentives to build a decentralized data marketplace for autonomous driving and Physical AI. Edge AI models on contributors' smartphones and VX360 dashcams analyze video in real time, strip personally identifiable information, and package structured events with blockchain-verified provenance for enterprise use. Valeo co-developed an open-source World Foundation Model using Natix data, and in July 2026 the University of Freiburg used contributor footage to release Orbis 2, its second autonomous driving foundation model. The NATIX token on Solana compensates contributors for validated data and gates enterprise access to real-world sensor datasets from 2.89 million drivers across 33 countries. Solana's throughput and low fees handle the micro-payment flows required for per-contributor incentives at scale. As of 2026, Natix has expanded from autonomous driving to robotics under a Decentralized Physical AI (DePAI) framework, applying the same infrastructure to developers of humanoid AI systems requiring grounded real-world sensor training data.

Extropy publishes original blockchain research through its Medium publication and Extropy Academy, covering topics including cross-chain MEV market structure, quantum cryptography threats to Ethereum, and EIP protocol analyses. A notable 2024 to 2025 study examined arbitrage and MEV markets across Ethereum, Solana, Optimism and Base, and StarkNet, producing detailed findings on how power-law concentration, per-trade economics, and infrastructure costs differ across ecosystems. The firm also produces a weekly Extropy Security Bytes digest documenting significant security incidents across Web3. The firm's research capabilities are reinforced by a team with backgrounds in cryptography, formal verification, and embedded systems, and by its Oxford base near academic research in these fields. Extropy's consulting services include zero-knowledge system design, domain-specific language research, and protocol-level security reviews. This combination of published analysis and applied research positions Extropy as an independent research voice covering both theoretical cryptographic advances and the applied security challenges facing live blockchain protocols.

Orbiter Finance uses zero-knowledge cryptography as a foundational layer for its cross-chain arbitration and dispute resolution system. The ZK-SPV (Zero-Knowledge Simple Payment Verification) component authenticates transaction existence, timing, and rule compliance across heterogeneous chains without requiring a full node, enabling trustless dispute resolution without relying on an oracle or centralized relayer. Beyond the bridge, Orbiter has developed the zkProver, which aggregates up to 128 ERC-20 user-operation signature verifications into a single ZK proof, reducing average gas consumption by roughly 58% compared to standard account-abstraction accounts. The Aggregated zkProver extends this into a recursive SNARK scheme allowing ZK rollups to participate in shared proof aggregation, launched in Q4 2024. The team also co-authored a zero-knowledge research paper with the Ethereum Foundation in 2024.

RISC Zero is a zero-knowledge computing platform that enables developers to generate cryptographic proofs for arbitrary Rust programs without requiring ZK mathematics expertise. The zkVM runs programs compiled to the RISC-V instruction set and produces zk-STARK proofs attesting to correct computation, which are then aggregated and converted to Groth16 SNARKs for on-chain use. R0VM 2.0, released in April 2025, cut Ethereum block proving time from around 35 minutes to 44 seconds and reduced proof costs roughly 5x. Boundless, a decentralized proof marketplace, launched Mainnet Beta in July 2025 on Base and processed 542.7 trillion RISC-V cycles across 399,000 fulfilled orders by mid-August 2025. Multi-chain support includes a Solana Groth16 verifier using native alt-bn254 system calls, audited by Veridise. Wormhole, Citrea, and Nethermind each integrate RISC Zero for cross-chain verification and validity proofs.

Velvet Capital's AI layer is built around a multi-agent system named Velvet Unicorn, comprising four specialized agents — CEO, Researcher, Analyst, and Trader — that coordinate to respond to natural-language queries and combine onchain metrics with social sentiment and technical analysis before preparing or executing trades. In August 2026, the platform released Velvet Flash 0.1, a 4-billion-parameter language model purpose-built for crypto trading tasks, intended to give Velvet tighter control over inference quality for DeFi-specific reasoning. Users can interact with the system using plain-language queries and receive structured research outputs or automated trade executions. This in-house model supplements the multi-agent framework and forms the backbone of the Velvet Unicorn AI infrastructure deployed across the terminal, VelvetX, and the VU Bot.

CrunchDAO is a decentralized AI research network that distributes machine learning challenges to 10,000+ engineers and 1,200+ PhDs across 100 countries, then aggregates their competing models into production-grade prediction feeds. Organizations submit forecasting problems with proprietary data secured inside Trusted Execution Environments, allowing participants to train models without accessing raw inputs. An automated scoring engine evaluates submissions against live data, ranks them on a real-time leaderboard, and deploys top models into client cloud infrastructure. Results include a 17% improvement in asset pricing accuracy for ADIA Lab and cancer research breakthroughs with the Broad Institute of MIT and Harvard. Solana serves as the settlement layer for real-time leaderboards and weekly USDC reward payouts directly to participants' wallets. CrunchDAO joined the Solana Incubator Cohort 2 in January 2025, deploying audited contracts for reward escrow and performance tracking. The CRNCH governance token gives participants voting rights over Coordinator approvals, emissions rates, and ecosystem grants.

Chakra Labs built its verifiable data warehouse with AI use cases at the center, offering on-chain provenance for web data so developers and researchers can verify the origin and integrity of training or inference inputs. The warehouse layer includes Model Context Protocol support, connecting Chakra data directly into AI chat interfaces such as Claude Desktop and ChatGPT via local MCP servers, allowing language models to query the warehouse during inference. By mid-2025 the company had shifted its public identity toward AI agent infrastructure, building deterministic reinforcement learning environments, high-fidelity trajectory datasets for computer-use agent training, and benchmarks such as MAGI. GitHub activity reflects this pivot with repositories for agent evaluation frameworks including Marina and Harbor, and mocked single-page applications for RL environments.

Dark Research is an AI lab building infrastructure for private AI agent operations on Solana, using Trusted Execution Environments and Anthropic's Model Context Protocol. The project treats AI agents as active on-chain participants — trading, managing wallets, executing smart contracts — and wraps their computation inside hardware-isolated enclaves to guarantee confidentiality even from node operators running the underlying servers. Applied research surfaces through DARK Forest, an AI simulation game where agents compete inside TEE-secured environments and serve as live benchmarks for agent capability, and through SPUTNIK, the project's first MCP-augmented competitive AI agent designed for incomplete-information games. Founded by Marginfi co-founder Edgar Pavlovsky and backed by MtnDAO, Dark Research positions itself at the intersection of AI capability research and blockchain-native execution environments.

Gradient Network advances AI and blockchain research through open protocols for distributed model inference and reinforcement learning on Solana. Its Echo framework uses a Dual Swarm Architecture to decouple RL inference from training, enabling large-scale runs across consumer hardware without requiring every device to support full GPU workloads. Published Echo results include a 7B math model outperforming Qwen2.5-32B and a 30B reasoning model surpassing DeepSeek-R1 on the Sokoban benchmark. Gradient raised $10 million in seed funding in June 2025 from Pantera Capital, Multicoin Capital, and HSG to develop Lattica, Parallax, and Echo — positioning the project against Bittensor, SingularityNET, and Gensyn. The team open-sourced Parallax in late 2025, contributing the distributed inference engine to the research community. Co-founder Eric Yang frames the mission as making intelligence a public good by routing AI workloads across decentralized hardware coordinated on-chain through Solana.

MycoDAO is building a multi-agent AI analysis platform on the elizaOS framework to process fungal genomic sequences and field observation data generated by its distributed citizen science network. The MycoApp functions as a gamified field identification tool allowing amateur and professional mycologists worldwide to submit verified fungal observations in exchange for $MYCO token rewards, producing a dataset at a scale no single institution could replicate independently. On-chain data anchoring through Bitcoin inscriptions creates permanent, timestamped records of genomic discoveries for provenance and priority-claim purposes, while Solana handles the high-frequency micropayment flows that reward data contributors in real time. The portable Lab-in-a-Box system — containing DNA sequencing equipment, sterile cultivation hoods, microscopy hardware, and automated cultivation chambers — provides the field hardware layer that feeds raw biological data into the AI analysis pipeline.

UpRock sits at the intersection of decentralized infrastructure and applied AI, using its contributor bandwidth network as the foundation for two distinct AI products. The Knowledge Acquisition Layer (KAL) orchestrates edge devices to execute browser-based crawls, then processes raw content through ML captioning, sentiment analysis, and LLM extraction to produce structured data ready for AI pipelines without additional cleaning. Businesses and developers access this synthesized intelligence through the AI Insight Exchange (AIX), a conversational dashboard that accepts natural-language queries and returns structured insights drawn from distributed crawls across global news sources covering five continents. The project roadmap extended this architecture into a consumer AI agent called Rocket. Unlike prompt-response chatbots, Rocket runs persistent background tasks monitoring prices, tracking topics, and following calendar events, surfacing results only when something actionable appears. A 2026 capability called Video Semantic Search enables natural-language queries over video archives, allowing users to locate spoken content or specific moments in footage without relying on title metadata. UpRock posits that blockchain-incentivized data collection can produce AI training and inference inputs that centralized crawlers cannot match in geographic diversity or authenticity.

WootzApp sits at the intersection of AI data collection and blockchain infrastructure by converting browser-level user activity into labeled training data for machine learning models. Every search query, page visit, and click interaction generates signals that post-training refinement pipelines consume, with collection integrated at the renderer level rather than through third-party scripts. Solana serves as the native blockchain backbone, handling crypto reward payouts to users who consent to contribute their browsing data to this marketplace. The project won second prize in the Consumer Apps track at the 2024 Solana Renaissance Hackathon, competing among more than 8,300 participants from 95 countries. This placed WootzApp within the first wave of consumer applications experimenting with tokenized AI data contribution on a public blockchain. The open-source codebase, published under AGPL v3, makes the collection methodology independently auditable, which is a meaningful trust signal for a product that positions user-generated browsing signals as the raw material for AI model improvement.

Leea Labs builds protocol-level infrastructure for deploying, orchestrating, and monetizing multi-agent AI systems on Solana. The project acts as a middleware layer connecting hardware providers, AI developers, and end users through a trust-minimized framework, with Solana smart contracts governing agent registration, payment escrow, workload distribution, and dispute resolution. The platform integrates with established AI frameworks — LangChain, LangGraph, CrewAI, AutoGen, and Swarm — letting teams adopt the infrastructure layer without rewriting existing agent logic. The LEEA token anchors the agentic economy by funding task-based agent payments through escrow, securing the validator network through staking, and granting governance rights over protocol parameters including fees and validator penalties.

OAK Research is a crypto data and research platform that produces long-form analysis, deep-dive reports, and market briefings covering both digital assets and traditional finance markets. The team, founded in 2024, includes former financial engineers, a DeFiLlama contributor, and editors from France's largest crypto media outlet, Cryptoast. The platform's most substantial research takes the form of year-end collaborative reports — its 2026 edition ran to more than 180 pages with 12 analysts and 20 contributors, produced alongside Castle Labs and Hazeflow. Its investigations examine protocol mechanics, governance proposals such as EIP-8363 and the Lean Ethereum proposal, and structural market questions including altcoin underperformance and the long-term sustainability of DeFi yields.

Agent Kitsune is Raposa Coffee Co.'s AI-powered commerce agent, built in partnership with Griffain, an agent infrastructure platform on Solana. It allows users to purchase Raposa Coffee products directly through the Griffain platform using an agent-to-agent trustless payment system that runs on Solana's payment rails, integrating with Raposa's existing Shopify backend. By routing transactions through Solana rather than traditional payment processors, Agent Kitsune eliminates the standard 2–3% credit card fees common in conventional e-commerce. The agent also learns customer preferences over time, functioning as both a transactional layer and an informational assistant for the brand. The longer-term roadmap for Agent Kitsune reaches beyond Raposa's own product catalog. The team has outlined plans to extend the same agentic commerce framework to other web2 platforms including WooCommerce, Square, and Wix, bringing additional brands onchain through Solana's payment infrastructure. This positions Agent Kitsune not just as a tool for Raposa customers but as a potential general-purpose commerce agent capable of connecting any retail brand to Solana's payment rails. The initiative is one of the more concrete demonstrations of how AI agents and blockchain payment infrastructure can combine to reduce friction in real-world consumer transactions.

AIxBlock operated as a Solana-based decentralized AI development platform from around 2023 to early 2025, offering a GPU compute marketplace, a model marketplace, and low-code AI workflow automation. The platform promised compute costs up to 90% lower than traditional cloud providers, and a Token Generation Event was announced for Q2 2025. A January 2025 partnership with node management platform Hivello broadened its distributed compute network, with the initiative backed by the European Union's innovation programme. The project drew on a seven-year track record delivering AI training data to clients including Oracle, AWS, Apple, and Chainalysis — covering more than 200,000 hours of annotation and audio work across 100+ languages — to bridge enterprise AI demand with decentralized compute supply on Solana. By mid-2025, AIxBlock had withdrawn its public token and marketplace infrastructure, repositioning as a self-hosted enterprise AI data platform with no active Solana presence.

Donut is designed as AI-native infrastructure for DeFi trading rather than a conventional interface with machine learning features added on top. The D0 agent runs a multi-model orchestration stack for routing, planning, and execution, paired with real-time critic loops and on-chain-grounded prompt benchmarks. The founding team draws from applied AI research backgrounds at Meta AI, TikTok AI, and Tencent AI, and the system is explicitly framed as a reinforcement learning environment for DeFi agents — treating market engagement as an AI reasoning problem that Solana's speed and composability are suited to support. The Agentic Loop processes signals from charts, on-chain data, whale activity, and funding rates simultaneously, then streams every reasoning step and data source back to the user in real time. Users interact with D0 across a web dashboard, Telegram bot, CLI, and voice input, issuing instructions in plain language and receiving transparent reasoning in return. Donut's $22 million in funding from investors including BITKRAFT, Makers Fund, and Sky9 Capital, alongside ecosystem partnerships with Solana, Sui, and Injective, positions it as a foundational layer in the emerging AI-agent trading stack on Solana.

3NS formalizes the intersection of AI agents and blockchain infrastructure through its open-source Agent Domain Protocol (ADP), published under MIT license at github.com/3nsai. The protocol defines a structured standard for attaching AI logic to blockchain domains and making that intelligence discoverable across networks, addressing what agent-to-agent trust and identity should look like on-chain. ADP is deliberately framework-agnostic, supporting implementations built with LangChain, Google GenKit, Autogen, CrewAI, Fetch.ai, Ollama, and OpenWebUI. The open-source TypeScript ADP implementation gives developers tools for building agent-enabled domains and connecting them cross-chain. Workflow automation integrates with N8N and ActivePieces for both no-code drag-and-drop and programmatic agent configurations. A WebRTC-based agent-to-agent communication layer is in private development, reflecting ongoing research into trustless machine coordination at the protocol layer.

Spot On Chain applies machine learning to blockchain data at scale, running proprietary AI models trained over three years to classify wallet types and identify capital flow patterns across 12 networks. The platform processes approximately 20,000 on-chain transactions per second through its data pipeline, translating raw ledger activity into behavioral signals that separate sophisticated traders from ordinary retail participants. Report generation leverages Google Gemini 1.5 Pro via Vertex AI, automating 90% of blockchain signal report production with each report completing in roughly five seconds. The Blockchain Personal AI Analyst component flags irregular on-chain events for individual users in real time. This applied AI approach positions Spot On Chain among machine intelligence tools designed to interpret blockchain data rather than merely surface it.

ZKP2P originated as an Ethereum Foundation Privacy and Scaling Explorations (PSE) research project, with its first protocol version built entirely on zero-knowledge proofs. V1 used zkEmail — ZK proofs of DKIM signatures embedded in payment confirmation emails — compiled using circom and halo2 circuits to verify that a fiat payment occurred without exposing full email contents to any third party. The attestation layer evolved through V2, which added proxy-TLS and TLSNotary (MPC-TLS) approaches, and into the current V3, which uses AWS Nitro Enclaves to produce EIP-712 signed payment attestations gated to the enclave specific code measurement. The open ZKP2P protocol retains its research lineage as a working demonstration of applying ZK and cryptographic attestation techniques to trust-minimized, on-chain fiat payment verification.

LYNC has expanded its developer infrastructure into agentic use cases, offering tools for AI agents to hold wallets, execute transactions autonomously, handle payment approvals, and interface with on-chain logic. This targets the growing category of autonomous on-chain agents, where AI systems need reliable programmatic access to blockchain infrastructure without human signing at each step. The tooling extends LYNC's embedded smart wallet layer to non-human principals across its supported networks. By combining gas abstraction, autonomous transaction execution, and MPC key management within a single SDK, LYNC positions its agentic tooling as production-ready infrastructure for AI-driven on-chain applications. The platform operates across 14+ blockchains including Solana and already powers 500+ integrated projects and 50,000+ user operations, making it one of the few cross-chain developer platforms to explicitly address autonomous AI agent wallet infrastructure.

RPS AI is an applied AI lab that pivoted in 2024 to focus on algorithmic market making on Solana, positioning itself where real-time data infrastructure and machine learning intersect. Its flagship product, UltraLiquid, uses a continuously learning model trained on historical Solana token launches to manage DEX liquidity positions autonomously. A standout transparency mechanism is the TEE-signed report produced after major risk events, providing a cryptographically attested record of all algorithmic actions using a Trusted Execution Environment signature. The AI layer extends to RPS AI's proprietary data infrastructure, which accesses liquidity data from deeper market sources with multiple geographic and provider fallbacks rather than standard Solana RPC endpoints. This enables near-real-time rebalancing at the speed of Solana's block times across approximately 50 million daily transactions. The data infrastructure is also offered as a standalone service for market makers, AI agents, and DeFi protocols needing high-uptime, low-latency on-chain data.

Polytrade's 2025 repositioning frames the company as the launchpad for AI-powered finance products, integrating artificial intelligence across three distinct product lines. Polytrade Finance uses AI to verify deal documentation and complete SME credit underwriting within minutes — compressing a process that traditionally requires days of manual review. Ledgerless, a live accounting product for UAE-based trading SMEs, employs AI-assisted invoice capture, double-entry ledger management, and VAT preparation, automating the documentation layer beneath trade finance. The AI thesis is that five years of RWA infrastructure — compliance rails, KYC pipelines, asset standards — become a powerful foundation when combined with AI-driven automation for the financial workflows sitting on top. The most forward-looking AI product is Edge, currently in development: an AI agent for trading tokenized stocks and real-world assets in a non-custodial setting, with integrated research tools and risk management features. Edge positions Polytrade in the emerging category of AI-driven investment agents operating directly over on-chain asset markets — a space where blockchain execution and AI decision-making converge. The project's roadmap also includes a peer-to-peer options market for hedging or speculating on tokenized assets, which could eventually serve as execution infrastructure for AI-driven hedging strategies. Institutional backing from Mastercard and the team's grounding in traditional trade finance compliance give this AI-finance convergence thesis a more operational foundation than most blockchain-AI projects.

NewMoney.ai builds its entire product around Newton, a conversational AI agent that executes blockchain transactions from inside messaging threads on WhatsApp, Telegram, Discord, SMS, and email. Newton interprets plain-language requests — such as "swap SOL for ETH" or "send 20 USDC to Maria" — and handles the corresponding on-chain execution without requiring users to interact with a conventional wallet interface, addressing crypto's UX barrier through natural language rather than UI redesign. The company also publishes a developer-facing SDK at caishen.tech enabling third parties to build custom AI agents on top of its infrastructure, positioning it as a potential platform layer beyond the consumer product. This dual model — consumer AI wallet plus developer toolkit — represents one of the more complete implementations of the AI-agent-native finance stack currently active on Solana.

Meme Money Markets is Moonit's fully automated AI pipeline that monitors 9GAG, a meme platform with over 200 million registered users, and mints new Solana tokens the moment a post crosses viral engagement thresholds. The system tracks likes, shares, and views in real time with no human decision-making in which content gets tokenized; there are no founders, pre-sales, or developer allocations on these launches. The AI triggers token creation algorithmically when a post crosses defined popularity signals. Announced on July 31, 2025, the 9GAG partnership produces tokens identifiable by a '9gag' suffix in their contract address, distinguishing them from standard Moonit Fun launches. Moonit has also published an Auto-Launch SDK enabling other content platforms to implement similar culture-to-crypto pipelines, extending the model beyond a single partner. The system is designed to let communities directly capture the financial value of their own virality, bypassing platform advertising as the sole beneficiary of viral engagement.

Motif AI sits at the intersection of artificial intelligence and blockchain-native finance, building a proprietary time-aware knowledge graph called Clarity to power its advisory services. Rather than relying on conventional retrieval-augmented generation, Clarity uses Contextual Edge Enrichment to give connections between financial entities temporal weight, surfacing how asset truths have evolved rather than only what is currently true. Clarity operates through three coordinated agent swarms: ingestion agents processing financial data and market signals, curator agents adding temporal context, and response agents synthesizing natural-language answers. This AI infrastructure powers Motif's full advisory suite, delivering personalized portfolio recommendations and market insights to both institutional white-label partners and retail investors through a consumer mobile application.

IoTeX positions itself as infrastructure for real-world AI, bridging verified physical-world data to AI inference pipelines. QuickSilver is an AI interaction layer that transforms raw machine data streams into structured, verifiable AI context usable by AI models in real time. The June 2026 Anti-Roadmap initiative introduced Trio, a multimodal foundational AI model designed to enable AI to understand and operate in the physical world using verified data streams from the device network. W3bstream zero-knowledge proofs for data integrity combine with QuickSilver to deliver cryptographically verified real-world inputs to AI models rather than synthetic data alone. ioID gives devices and AI agents verifiable on-chain identities, enabling trusted machine-to-model interaction. The founding team includes cryptography researchers with experience at Google, Uber, and Facebook.

Nodepay's Signal Engine is an autonomous AI system that continuously ingests public data from social feeds, crypto news platforms, market data providers, and on-chain activity to produce structured market intelligence. The engine employs a Web3-specialized multi-label classifier to categorize signals by narrative — BTC, L1, AI, DePIN — and an embedding model that converts signals into numerical vectors for semantic clustering beyond simple keyword matching. An integrated LLM then synthesizes clustered signals into human-readable narratives explaining price movements, emerging attention, and narrative formation. These AI-generated outputs power Node Pulse and Node Search, giving traders and analysts natural language access to a continuously updated intelligence archive.

ChainGPT is a full-stack AI ecosystem built for Web3, offering specialized tools including a Web3 AI Chatbot trained on crypto and blockchain data with real-time on-chain access, an AI Trading Assistant covering technical analysis across 5,000+ tokens, and an AI-Generated News Feed. Founded in 2022 by Ilan Rakhmanov, the platform serves 550,000+ monthly active users. ChainGPT's AIVM is a Layer-1 blockchain built on the Cosmos SDK for decentralized AI execution, featuring dual-path execution where simple models run fully on-chain while complex workloads run off-chain with cryptographic proofs. The CGPT token launched as an SPL token on Solana in May 2025, with integrations spanning Metaplex, MagicEden, LayerZero, and Stargate for cross-chain interoperability.

Walrus is built around RedStuff, a proprietary two-dimensional erasure coding scheme developed by Mysten Labs that encodes stored blobs into slivers distributed across independent storage nodes. The system achieves Byzantine fault tolerance guarantees, ensuring full data recovery even when two-thirds of storage nodes go offline or act maliciously, while requiring only 4–5× data replication overhead. All data stored on Walrus carries cryptographic proof of correctness, allowing anyone to independently verify data integrity without trusting the storage provider. A December 2025 technical paper on arXiv describes a hybrid Trusted Execution Environment (TEE) and MPC architecture where TEE nodes run inside the validator network to strengthen cryptographic security guarantees beyond what standard MPC provides, addressing the challenge of secure private key management without centralized control.

AiSol integrates AI directly into the Solana liquid staking experience through AiMate, an AI-powered portfolio manager accessible at aimate.aisol.fi that provides AI-assisted tools for managing Solana-based DeFi positions. The project also channels a share of staking rewards into a fund targeting early-stage AI projects building within the Solana ecosystem, connecting passive staking capital to active AI builders. Developed by AiNodes-Tech, a blockchain infrastructure company holding official Circle partner status and infrastructure grants from Google Cloud, DigitalOcean, and OVH, AiSol's roadmap includes an AI Infrastructure Layer planned for early 2026 providing dedicated blockspace, low-latency RPC access, and compute resources to blockchain projects. This positions AiSol as one of the few Solana platforms combining AI tooling at the user level with AI ecosystem funding and infrastructure at the protocol level.

Oraichain is the world's first AI-powered Layer 1 blockchain oracle, built to resolve the three structural barriers preventing AI integration with smart contracts: determinism requirements, language incompatibility, and on-chain storage constraints. It operates as middleware that validates AI service outputs before writing results to any connected chain, letting smart contracts on Solana and elsewhere consume AI computation without compromising on-chain guarantees. The AI oracle mechanism is driven by test-case-based validation — requesters submit test cases alongside each computation request, and validators independently execute those cases before accepting results, with slashing penalties for poor output quality. Developers publish AI algorithms to an AI Marketplace where dApps can discover and integrate them, while the Data Hub supports decentralized data creation. Ecosystem products including Thesis.io (60 billion AI tokens processed by H1 2025), Agents.land (over 1,000 on-chain AI agents deployed), and LFG!!! (AI-driven meme-coin trading signals on Solana) illustrate the range of AI-blockchain applications the network enables.

PotBot's Blockchain Orchestration Tool (BOT) demonstrates AI-assisted on-chain capital management with enforced human oversight. The BOT reads vault state and market prices to generate swap proposals, routing execution through Jupiter only after an on-chain governance threshold is met. An MCP server at @potbot/[email protected] lets external AI systems including ChatGPT and Claude connect, propose trades, and receive group consent before any capital moves. Every BOT action is paired with a signed reason string stored on-chain alongside the governance vote, creating an auditable trail of AI recommendations and approvals. Delegation is per-vault and revocable at any time, ensuring no autonomous system can unilaterally control pooled assets. PotBot offers a practical reference implementation for human-AI collaborative treasury management on Solana, with StrategyTrigger automation and inter-vault POT Duels on the roadmap.

CyreneAI integrates artificial intelligence directly into on-chain workflow automation through its Flows product, which powers AI-driven trading agents capable of monitoring social signals and executing decentralized transactions autonomously. The flagship Trading Agent reads Telegram channels for token signals, intelligently validates contract addresses, and executes on-chain trades with configurable risk parameters — demonstrating practical AI-blockchain convergence on Solana. This moves beyond AI tools that merely analyze blockchain data to systems that act on it in real time. The Genesis Accelerator program explicitly scopes AI and automation projects as priority applicants, signaling CyreneAI's strategic intent to serve as infrastructure for AI-native on-chain applications. By combining visual workflow tooling with native integrations into Jupiter, Raydium, and Pyth, CyreneAI gives AI agents the building blocks to participate in Solana DeFi without custom contract development. The result is a platform where AI-generated signals translate directly into on-chain economic activity.