Open-source decentralized search layer

    Desearch is the live-context layer for AI agents.

    Real-time web data, open ranking, and competitive validation. Built on decentralized incentives, exposed through APIs developers can ship today.

    Explore the APIBehind the tech
    Subnet 22
    Bittensor network for decentralized search
    10+
    source families across web, social, docs, and research
    1 API
    search, crawl, social context, and cited AI answers

    Network map

    From raw public data to validated context.

    Public sources

    01

    web, X, docs, research

    Provider competition

    02

    freshness and coverage

    Validator scoring

    03

    quality and relevance

    API output

    citations
    ranked links
    structured JSON
    cost metadata

    Fresh data should be agent-ready

    Agents need current context, source links, snippets, and predictable response contracts instead of brittle scraping stacks.

    Ranking should be measurable

    Independent providers compete on quality while validators score freshness, relevance, and reliability in the open.

    Powered by Bittensor

    Built on decentralized incentives.

    Desearch runs on Bittensor, a decentralized AI coordination network where independent participants compete to provide useful work.

    Miners collect fresh public data. Validators evaluate quality. Token incentives reward outputs that are accurate, relevant, and useful for developers.

    TaostatsCoinGecko
    Taostats Dashboard showing Desearch subnet data

    How Desearch works

    From raw data to validated intelligence.

    Collect live data visualization
    Miners / 01

    Collect live data

    Independent data providers crawl public sources and submit fresh web and social context.

    See miners
    Score quality visualization
    Validators / 02

    Score quality

    Validators evaluate relevance, accuracy, and freshness so the strongest outputs move up.

    See validators
    Ship context to agents visualization
    API / 03

    Ship context to agents

    Developers call one API for search, crawl, X, and AI summaries with citations and cost metadata.

    View API docs

    Trusted by AI innovators worldwide

    Built for teams that need fresh web, social, crawl, and cited AI context in production workflows.

    Bittensor iconbittensor
    AWS logo
    Lovable logo
    Cursor logo
    Notion logo
    LIUM logo
    Chutes
    TaoMarketCap

    Build or contribute

    Use the API or help operate the network.

    Desearch supports both sides of the market: builders who need live context and providers who contribute data quality.

    For data providers

    Join as a miner and contribute to the intelligence layer. Crawl public sources, deliver quality data, and earn rewards.

    Get started

    For developers

    Integrate decentralized search into your agents using the unified API, SDKs, docs, and hosted console.

    Open console
    Community proof

    Builders are already using Desearch for live agent context

    Real comments from users wiring Desearch into research, crypto analysis, and agent workflows.

    View all voices
    ExponentialReed
    ExponentialReed
    @ExponentialReed

    Totally agree with this, i just connected mine to Bittensor $TAO subnet 22 @desearch_ai with an API end point for X data to help me with subnet analysis and the results are getting where they should be. This could be used by anyone in crypto doing reporting and research and is an incredible opportunity based on pricing.

    Tweet attachment
    53Reply
    Open
    CryptoZPunisher
    CryptoZPunisher
    @CryptoZPunisher

    SN22 is becoming an intelligence layer for agents. And behind agents… come predictions, then portfolio management. Revenue + aligned incentives. The market probably hasn't fully priced this yet.

    Tweet attachment
    43Reply
    Open
    Numinous AI
    Numinous AI
    @numinous_ai

    The forecasting environments just got richer. Starting this Thursday, miners can choose to take on the cost of news retrieval via @desearch_ai The cost limit will move from $0.02 to $0.1 - making our environments ready for heavy forecasting tasks. #SN6 #TAO

    53Reply
    Open
    vaNlabs
    vaNlabs
    @vaNlabs

    // Desearch integration complete crustty now has full social sentiment guidance and its integrated (weighted) in his subnet valuations. All thanks to Bittensor SN22 @desearch_ai Desearch vs. native X API below. I gave him both and asked to research @HermesSubnet, no bias.

    Tweet attachment
    28Reply
    Open
    Builder community

    Get closer to the search network.

    Join the channels that help you ship: API examples, hard-query feedback, product release notes, and conversations with the team building Desearch and Subnet 22.

    Join DiscordAPI quickstart

    Discord

    Implementation help, query feedback, and direct product discussion.

    Join Discord

    X (Twitter)

    Product releases, benchmark notes, and Bittensor ecosystem updates.

    Follow on X

    Telegram

    Release notes and community announcements in a lightweight channel.

    Open Telegram

    LinkedIn

    Company updates, partnerships, hiring, and product milestones.

    View LinkedIn

    Give your agents live context.

    Start with one API key, test real requests in the playground, and ship search-augmented agents with current web and social data.

    Get API KeyView Docs

    Need something custom?

    We work with enterprises and teams that need custom integrations, dedicated support, or specific data requirements.

    Contact us