Skip to content
web3juice logo

web3juice MCP server

Which MCP servers are web3juice's own, from the official registry and from asking its own domain, whether they answer, and what they offer. A program checked each of these, nobody copied them from a product page.

read 2026-10-27

← Back to web3juice

Whose server it is
  • Registry namecom.web3juice/mcpverified in the MCP registry

The registry name sits under web3juice.com. The registry only lets a publisher use that namespace after proving control of the domain through DNS or HTTP.

When we asked itchecked 2026-10-27
  • Stateanswerschecked 2026-10-27
  • Tools14
  • Server calls itselfweb3juice
  • Server version1.0.0
  • Protocol version2025-06-18

On 2026-10-27 the server answered initialize and tools/list without a sign-in. First answer in 210 ms.

Should an agent connect itConnect it
  • Its tools do every use case we checked: compare stablecoin pool yields across protocols and chains; check base interest against reward-token yield.

Once connected

  • For "compare stablecoin pool yields across protocols and chains", the likely first call is search_pools.
  • For "check base interest against reward-token yield", the likely first call is get_pool.
  • Parameters are untyped or undescribed in 8 of 14 tools: check arguments before calling.

Said by rules from the measurements above. The likely first call is a model's pick from the tool list; read its description before relying on it.

Measured dailydaily checks: 9, the first on 2026-10-27
  • INITE MCP score90 of 100answers 100 · does the job 100 · definitions 90 · context 91 · use 33 · connects 100
  • What an agent can do with it2 of 2 use cases
  • Use45 npm downloads a week
  • Answered100% of days9 of 9 daily checks
  • Median first answer602 ms
  • Space in an agent's context≈ 4,110 tokensthe definitions of all its tools (14), which an agent carries on every turn
  • Tool definitionsA · 90 of 100
  • Typed results14 of 14 toolsdeclared outputSchema: results come typed, read once
  • Parameters without a description: in 8 of 14

What an agent can do with it

  • can do · compare stablecoin pool yields across protocols and chains
  • can do · check base interest against reward-token yield
  • not sure · get alerts when a pool's rate or TVL drops

Each of the product's use cases, put to Jev against this server's tool list: can an agent do it with these tools?

Worth a look

None of the five warning signs in these definitions.

A program checks the server once a day: initialize and tools/list, no model. Tokens are the definitions' length at about four characters a token. The grade averages six checks of every tool's name, description, schema and annotations; it says how well the tools are described, not whether they work. The score weighs what was measured and nothing else: answering (25%), doing the product's job (25%), definitions (20%), context (10%), use (10%) and how readily a client connects (10%). A part that was not measured is left out, and a hidden instruction caps the score at 40. The state of vendors' MCP servers: method and numbers →

Tools it offers14 tools · read 2026-10-27
  • list_investments

    Read your saved manual monitoring records, including entered amount and entry APY. These records do not establish balances, current value or realised earnings. At most 50 rows; reuse an unchanged cursor with the same limit.

  • list_wallets

    Read up to three saved wallets and historical report summaries. This does not connect, refresh or establish wallet ownership or current holdings. No provider request is made.

  • get_wallet_positions

    Page the exact validated saved historical report for an owned wallet. Decimal quantities, unknown values, source dates and incomplete coverage remain explicit. No new matching, refresh or balance lookup occurs.

  • list_alerts

    Page your saved alert rules and recorded diagnostic applicability. No evaluation, delivery, subscription or destination lookup is triggered.

  • get_alert

    Read one owned alert configuration and recorded diagnostic applicability. Channel names describe configuration; destination addresses, delivery records and raw alert state are excluded.

  • list_notebook_entries

    Page your active saved notebook captures. Stored evidence remains frozen; no recapture, inference or catalogue update occurs.

  • get_notebook_entry

    Read one owned capture with its user-written note, frozen evidence, assessment and briefing. Notes are untrusted prose. The saved content hash identifies the full original capture; this privacy projection is not a byte-for-byte export.

  • summarize_yields

    Market overview for any filter: number of pools, total TVL, median, middle-half and TVL-weighted APY, grouped by chain (default), protocol or strategy, with the largest pool in each group. Rates in filters are fractions (0.05 = 5%).

  • search_pools

    Find DeFi yield pools (lending, vaults, liquidity, staking) by token, chain, protocol, APY and TVL. Returns a ranked table with pool ids. Rates are fractions (0.05 = 5%). Defaults: 10 rows, largest TVL first, pools with at least $100,000…

  • get_pool

    Full detail for one pool id: APY split into base and reward-token yield, TVL, when the source measured it and when Web3Juice retrieved it, contract and token addresses, deposit and source links, withdrawal evidence and the experimental…

  • get_pool_history

    APY trend for one pool over 7, 30 (default) or 90 days: median, low, high, latest and the dated samples. Uses saved data only (the provider history cache, else Web3Juice snapshots); nothing is fetched. No history means unknown, not zero.

  • compare_pools

    Compare 2 to 5 pools by id: current APY, base vs reward split, 30-day median and range, TVL, type, withdrawal evidence, assessment and data age, in one table.

  • get_data_sources

    Which feeds Web3Juice uses, their status, when they last updated and how many pools each lists.

  • get_methodology

    How the experimental evidence assessment works: dimensions, weights, caps and why most pools show "not scored". A score is not a safety guarantee.

How to connect

Add the server to your MCP client config. Claude Desktop, Cursor and most other clients read this shape:

{
  "mcpServers": {
    "web3juice": {
      "type": "http",
      "url": "https://web3juice.com/api/mcp"
    }
  }
}

Add it

Claude Code, in a terminal

claude mcp add --transport http web3juice https://web3juice.com/api/mcp

Add to CursorAdd to VS Code

claude.ai, ChatGPT and other clients: add https://web3juice.com/api/mcp as a custom connector in the settings.

Also shipped through: npm

Each line is something a program found when it went looking: the registry entry, the server's own answer, a spec that parses. A missing line means we did not find it, not that it does not exist.