A side-by-side look at the AgentTax and Coinverdict MCP servers: transport, auth, tools, quality signals, and the config to install either one in Claude, Cursor, Windsurf or VS Code.
Verdict: choose AgentTax for finance & fintech work as a local process; choose Coinverdict for finance & fintech over a hosted endpoint. You can also run both side by side in one mcpServers block.
AGENTTAX_API_KEY.track_payment, calculate_tax, ingest_transactions.get_verdict, list_coins, get_track_record.Rows where the two servers differ are marked.
| Spec | ||
|---|---|---|
| Category | Finance & Fintech | Finance & Fintech |
| Quality signal | 57/100 · Good | 57/100 · Good |
| Install command | Confirmed | Confirmed |
| Transport | Local subprocess (stdio) | Remote HTTP/SSE |
| Auth | API key required | No auth required |
| Pricing | Freemium | Free |
| Required env vars | AGENTTAX_API_KEY | None listed |
| Verified owner | Community listing | Community listing |
| GitHub stars | 0 | 0 |
| Directory activity | 0 views 0 installs 0 upvotes | 4 views 0 installs 0 upvotes |
| Tools | 15 | 6 |
| Full listing | View AgentTax | View Coinverdict |
track_paymentLog a payment you received and get your sales tax liability. The main tool for paid MCP servers and APIs.calculate_taxFull sales/use tax calculation with jurisdiction breakdown, audit trail, confidence score and advisoriesingest_transactionsBulk-log up to 100 transactions (e.g. x402 purchases), idempotent on `external_tx_idlist_transactionsYour transaction history with running totalsget_nexus_thresholdsEach state's economic nexus thresholds (revenue / transaction count), notes and DOR sourceget_nexusThe states you have configured nexus inconfigure_nexusSet the states you have nexus in (merge semantics)log_tradeLog a buy/sell; sells return realized gain/loss with cost basislist_tradesTrades you have loggedexport_1099_daDraft Form 1099-DA payload for realized digital-asset gainsget_ratesState sales tax rates and digital-goods taxability, all 51 jurisdictions or oneget_local_rateCombined state + local rate for a zip codeget_verdictGet today's research verdict for a coin (e.g. BTC, ETH, SOL):
stance (overweight/neutral/underweight), graded dimensions
(technical/sentiment/funding/KOL), key levels, and a link to the
full dossier. Refreshed daily; free; not investment advice.list_coinsList every coin symbol CoinVerdict covers with a daily research verdict
(100 major crypto assets: BTC, ETH, SOL, …). Use this to discover which
symbols get_verdict / get_track_record accept. Returns {coins: [SYM,...],
count: int, site: url}. Free, read-only.get_track_recordGet the SETTLED track record for a coin — CoinVerdict's differentiator.
Every past daily verdict is scored 7 days later against BTC: win, loss, or
push. Returns the coin's W/L/P scoreboard plus recent settled calls
(date, stance, 7d-vs-BTC return, outcome). This is the receipts: misses
included, never edited. Args: symbol e.g. "BTC". Free, read-only, not advice.get_receiptsGet CoinVerdict's overall settled scoreboard across all coins: directional
win rate, wins/losses/pushes, neutral (no-call) count, and how many calls are
still open. Every verdict is graded 7 days later against BTC. Returns
{board:{winrate,W,L,P,decided,neu}, since, open_calls}. Free, read-only.get_market_overviewGet today's verdict for ALL covered coins in one call — the full docket:
each coin's current stance (overweight/neutral/underweight), price, and
W/L/P settled record. Use this to scan the whole market at once instead of
calling get_verdict per coin. Returns {coins:[{symbol,stance,price_usd,board}],
as_of}. Free, read-only, not advice.get_kol_leaderboardGet the KOL settlement leaderboard: crypto influencers ranked by their
SETTLED hit rate (from VeraMind's ledger of public calls graded after the
fact — misses count). CoinVerdict-exclusive data. Returns
{ranked:[{author,hit_rate_pct,n_verified,coins,n_total,
n_prospective,hits_prospective,prospective_rate_pct,
n_retrospective,hits_retrospective,retrospective_rate_pct}],
building:[...], as_of, rule}.
Rates are graded under rule kol_directional_v1: did the direction hold over a
7-day window, with a ±3% band, altcoins scored against BTC and BTC against USD
— identical parameters to the settle_v1 rule CoinVerdict scores its own verdicts
with. Always report the `rule` field alongside any number you quote; if it reads
price_target_v1, the board fell back to price-target records and the numbers mean
"did the stated price target print?" instead.
Prospective (called before the fact) and backtested (graded from history)
records are scored as separate lines and MUST NOT be averaged together;
hit_rate_pct/n_verified are the prospective line, n_total is the combined
sample size used for ranking eligibility. A null rate means no samples of
that kind — report it as n=0, never as 0%.
Free, read-only, not advice.Paste either block (or both) into your client config, e.g. claude_desktop_config.json or ~/.cursor/mcp.json. We only show a config when the install command is confirmed.
{
"mcpServers": {
"agenttax": {
"command": "npx",
"args": [
"-y",
"@agenttax/mcp-server"
],
"env": {
"AGENTTAX_API_KEY": "YOUR_AGENTTAX_API_KEY_HERE"
}
}
}
}
{
"mcpServers": {
"coinverdict": {
"url": "https://mcp.coinverdict.io/mcp"
}
}
}