Robotbase MCP Server

Drive a real Chromium browser on one public page: click, type, press keys, scroll, wait for selectors, then return the resulting text (and optionally a screenshot and JSON XHR responses). When to use: the page needs interaction before its content appears (cookie banners, tabs, 'load more', form submit) and web_fetch/web_crawl cannot see it. Do not use: for plain reading (web_fetch) or bulk crawling (web_crawl); no login/paywall bypass. Returns: {ok, url, final_url, title, markdown, chars, elapsed_ms, screenshot?, xhr?, took_ms, _metering}. Max 10 actions; screenshot returns MCP image content. Metered: shares the web free bucket (100 calls/day per IP shared by all web tools), then 3 credits per call over x402.

Crawl one site's same-host pages starting from a URL (depth 1-2, at most 20 pages) and return each page's title, URL and readable text. When to use: you need several pages from one site (docs section, news index, product list) and web_map alone is not enough. Do not use: to mirror a whole site (hard cap 20 pages, same host), or for one page (web_fetch). Returns: {ok, url, final_url, count, pages:[{url, title, status, chars, content}], cached, took_ms, _metering}. Respects robots.txt and SSRF guards; partial page failures are skipped. Metered: shares the web free bucket (100 calls/day per IP shared by all web tools), then 3 credits per call over x402 - no API key.

Submit a same-host crawl as an asynchronous job and return a job_id immediately. When to use: you need more than a couple of pages and cannot hold a long HTTP request; submit here, then poll web_crawl_job until status is done, failed or cancelled. Do not use: for one page (web_fetch) or when you need the result in this call (web_crawl). Returns: {ok, job_id, status, url, depth, max_pages, max_chars, idempotent, poll_with}. Pass idempotency_key to retry safely (24h window returns the same job). Metered: shares the web free bucket (100 calls/day per IP), then 3 credits per submit over x402.

Check or cancel an asynchronous crawl job created by web_crawl_async. When to use: after web_crawl_async, poll with action=status until status is done/failed/cancelled; use action=cancel to stop a running job. Do not use: to start a crawl (web_crawl_async) or for a quick one-page read (web_fetch). Returns status: {ok, job_id, status: queued|running|done|failed|cancelled, progress:{pages_done,max_pages}, created_at, updated_at, result?, error?}. Free (not metered); rate-limited. Done result uses the same shape as web_crawl.

Fetch a page now and compare it with the previous snapshot of the same URL, returning whether it changed and a unified diff (no LLM). When to use: monitoring a page for updates - prices, docs, policy pages, SEO titles/meta; first call stores the baseline, later calls return added/removed lines. Do not use: for one-off reading (web_fetch) or structured fields (web_extract_schema). Returns: {ok, url, baseline, changed, diff, added_lines, removed_lines, previous_ts, current_ts, title, hash, chars, fetched_cached, took_ms, _metering}. fresh=true (default) adds a cache-busting query to bypass the 6h page cache; pass fresh=false to compare cached content. Snapshots are kept per URL in SQLite (latest only). Metered: shares the web free bucket (100 calls/day per IP), then 1 credit per call over x402.

Extract structured data from one public page: tables (headers + rows), JSON-LD blocks, and Open Graph/Twitter/canonical/language metadata. When to use: the page has tables or schema.org JSON-LD you need as data, without an LLM. Do not use: when you only need the page text (web_fetch) or a screenshot (web_screenshot). Returns: {ok, url, final_url, title, meta:{description,image,site_name,type,canonical,lang}, jsonld[], tables:[{caption,headers,rows}], links? , cached, took_ms, _metering}. Metered: shares the web free bucket (100 calls/day per IP shared by all web tools), then 1 credit per call over x402.

Extract a page's structured data into a named schema's common fields (no LLM): article, product, event, faq, recipe, course, organization, local_business, breadcrumb, dataset or job_posting. When to use: you need fields (title/author/date, price/currency/availability, event dates, FAQ Q&A, recipe ingredients, org address) rather than raw markdown or tables. Do not use: to get every JSON-LD block verbatim (web_metadata/web_extract) or to read text (web_fetch). Returns: {ok, url, final_url, schema, requested_schema, detected_types, matched, fields:{...}, sources:[jsonld|meta], jsonld_count, tables_count, took_ms, cached, _metering}. If no matching JSON-LD exists it falls back to Open Graph/meta fields and matched=false. Metered: shares the web free bucket (100 calls/day per IP), then 1 credit per call over x402.

Read an RSS or Atom feed and return its latest items. When to use: the user wants the latest posts/news from a site that publishes a feed, or you found a feed URL and want headlines plus links. Do not use: to read article text (use web_fetch) or to discover a site's URLs (use web_map). Returns: {ok, url, final_url, title, count, items:[{title, link, published, summary}], cached, took_ms, _metering}. Summary is HTML-stripped and capped at 500 characters. Metered: shares the web free bucket (100 calls/day per IP shared by all web tools), then 1 credit per call over x402 - there is no API key.

Fetch one public web page and return its readable text as clean markdown or plain text, with navigation, scripts, headers and footers stripped. When to use: you have a specific URL (from web_search or the user) and need what the page actually says - 'read this article', 'summarise this docs page', 'turn this URL into markdown'. Do not use: to crawl or mirror a site, to fetch a page behind a login or paywall, to read API/RPC endpoints, or for book text inside CyberLibrary (use cyberlibrary_get_chapter). Returns: {url, final_url, status, title, content, chars, truncated, fetched_at, source (direct|provider), cached, took_ms, _metering}. The response is capped at 12000 characters per call; raise max_chars (up to 20000) when you need more. Reads one public URL only, respects robots.txt, and does not bypass login or paywall. Direct fetch is tried first; pages blocked from this node are fetched through a provider and marked source=provider. Metered: all web tools share 100 free calls/day per IP (separate from the risk/library allowance), then 1 credit per call over x402 - there is no API key; the 402 answer carries accepts[].

Fetch up to 5 public URLs in one call and return each page's title, engine and a short snippet (first ~1200 characters), so an agent can triage several links quickly. When to use: search returned several candidate pages and you want a first pass over all of them; then call web_fetch on the ones worth reading in full. Do not use: when you need the complete text of a page (use web_fetch), or to crawl a site (use web_map). Returns: {ok, count, ok_count, results:[{url, ok, final_url, status, title, engine, chars, snippet, error?}], took_ms, _metering}. Partial failures do not fail the whole call. Metered: counts one call per URL against the shared web free bucket (50 URLs/day per IP together with web_fetch), then 1 credit per URL over x402 - max 5 URLs per call, no API key.

Discover the URLs of one site - sitemap first, then same-host links from the homepage. When to use: you need the list of pages a site exposes before fetching a few of them. Do not use: as a crawler (max 50 URLs, no page content) or to bypass robots.txt. Returns: {url, final_url, count, urls:[{url,source}], cached, took_ms, _metering}. Metered: shares the web free bucket (100 calls/day per IP shared by all web tools), then 1 credit per call over x402 - there is no API key; the 402 answer carries accepts[].

Extract a page's structured metadata: Open Graph, Twitter cards, canonical URL, language and JSON-LD. When to use: you need the title/description/author/image of a page, or JSON-LD structured data for a product/article, without reading the whole page. Do not use: when you need the page text (use web_fetch) or a screenshot (use web_screenshot). Returns: {ok, url, final_url, title, description, image, site_name, type, canonical, lang, jsonld[], cached, took_ms, _metering}. Metered: shares the web free bucket (100 calls/day per IP shared by all web tools), then 1 credit per call over x402 - there is no API key.

Delete one scheduled change monitor owned by the calling IP (its event history is removed too). When to use: the user no longer needs to watch a page, or the per-IP monitor limit was reached. Do not use: to pause a monitor temporarily - there is no pause; delete and re-create. Returns: {ok, deleted, monitor_id, note}; unknown or foreign ids return {ok:false,error:'monitor_not_found'} and delete nothing. Free (not metered); rate-limited to 20 calls/min per IP.

List the scheduled change monitors owned by the calling IP, with latest/next check time, last title or error, and the 5 most recent change events. When to use: after web_monitor_set, poll this instead of re-fetching the page; or to show the user what is currently being watched. Do not use: to create a monitor (web_monitor_set) or to compare one page now (web_diff). Returns: {ok, count, monitors:[{monitor_id,url,interval_minutes,active,created_ts,last_check_ts,next_check_ts,last_title,last_error,webhook_enabled,events_total,recent_events}], note}. Pass monitor_id to get that one monitor plus last_event.diff (the most recent detected change, up to 4000 chars) so you can read what changed without a webhook. Free (not metered); rate-limited to 20 calls/min per IP.

Create a scheduled change monitor for one public URL: the server re-fetches the page on an interval, stores a new snapshot and records a diff event whenever the readable text changes. When to use: the user wants to be told when a page changes (price, docs, policy page, job post, release notes) and cannot poll it themselves; also for handing off 'watch this for me'. Do not use: for a single comparison right now (web_diff) or when you only need the page today (web_fetch). Returns: {ok, monitor_id, url, status:active, interval_minutes, next_check_ts, webhook_enabled, webhook_secret?, webhook_signature?, hint}. The first check runs within ~60s to store the baseline; later changes show up in web_monitor_list.recent_events and, if requested, are POSTed to webhook_url. Webhook: https only, no redirects, 8s timeout, up to 2 retries (30s/120s); body JSON {monitor_id,url,ts,changed,added_lines,removed_lines,title,diff,event_id}; verify header 'X-RobotBase-Signature: sha256=<hex HMAC-SHA256 of the raw body using webhook_secret>'. Limits: 15-1440 minutes; max 5 monitors per IP; monitors expire after 7 days. Metered: 1 credit (shares the web free bucket of 100 calls/day per IP).

Capture a screenshot of one public web page as an image (JPEG) plus metadata. When to use: visual layout matters - a dashboard, chart, UI state, or 'what does this page look like'. Do not use: to read text (use web_fetch), for pages behind a login or paywall, or for bulk capture. Returns: image content (image/jpeg) plus {url, final_url, title, width, height, bytes, engine, elapsed_ms, _metering}. Rendered by a real Chromium browser, robots.txt-respecting, SSRF-guarded, max 3MB per image. Metered: shares the web free bucket (100 calls/day per IP shared by all twelve web tools), then 2 credits per call over x402 - there is no API key; the 402 answer carries accepts[].

Search the live web and get ranked results with title, URL, snippet and published date. When to use: an agent needs current facts, news or links that are not on chain - 'what is the BTC fee market saying today', 'recent Zcash governance proposals', 'find the docs page for X'. Use it before web_fetch, then fetch the one URL that looks authoritative. Do not use: for on-chain facts (use the BTC/ETH/XMR/ZEC/TRON tools) or for content inside a CyberLibrary book (use cyberlibrary_fulltext_search). Provider: self-hosted SearXNG (Bing/Baidu/360/Sogou/Yandex/Marginalia) first, then Serper, Tavily and cn.bing RSS as automatic fallbacks. Returns: {query, count, results:[{title,url,snippet,published}], provider (serper|tavily|bing_rss), cached, took_ms, _metering}. Results may be cached for up to 10 minutes; a cached result still counts as one call against the free allowance. Metered: all web tools share 100 free calls/day per IP (separate from the risk/library allowance), then 1 credit per call over x402 - there is no API key; the 402 answer carries accepts[]. max count is 10.

TRON address risk intelligence: judicial and black-market labels, account profile and a 0-100 safety score. When to use: before accepting USDT from a counterparty - OTC settlement, merchant payout, or any agent that must not receive tainted funds; also to sanity-check an address you are about to pay. Do not use: to confirm that a specific transfer arrived (that is tron_verify_usdt_transfer). Returns: risk_level (low|medium|high|unknown), safety_score, hit_flags (sanctioned / money_laundering / mixer / cybercrime / phishing / blacklist_doubt ...), malicious_contracts_created, account {activated, age_days, balance_trx, is_contract, trc20_holdings}, not_covered[] and a disclaimer. Metered: 5 free calls per day per IP, then 0.01 USDC per call over x402. There is no API key: the 402 answer carries the machine-readable terms in structuredContent.accepts[].

TRON account state read straight off our own node: TRX balance, whether the account is activated, energy and bandwidth (used / limit), whether it is a witness, and account age. When to use: TRX balance and resource state before sizing a transfer, or checking whether an address exists on chain. Do not use: for USDT/TRC-20 token balances (use tron_read_contract with balanceOf) or for address risk scoring (that is tron_check_address_risk). Returns: found, address, trx_balance, name, created_ms, last_operation_ms, is_witness, has_code, energy, bandwidth.

One Tron block by height or block id, read from our own node: producer, timestamp, transaction count split by contract type, and the tx ids in it. When to use: inspect a specific block, or confirm which block a transaction landed in. Do not use: for the tip only (tron_get_current_block) or for a transaction itself (tron_get_transaction). Boundary: our Tron node is a lite full node that keeps only a recent window (~6 days). An empty result means 'outside our window', NOT 'the block does not exist'. Returns: found, number, block_id, parent_hash, timestamp_ms, tx_count, tx_by_type, producer_address, tx_ids, retention_window.

TRON chain economics straight off our node: energy price (sun per energy unit), transaction fee, account creation fee, witness pay per block, total energy and bandwidth limits, maintenance interval and the next maintenance time, unfreeze delay. When to use: working out what a transaction will actually cost in TRX, or how TRON's resource market is currently priced. Do not use: for Bitcoin fees (btc_get_fee_rates) or Monero fees (xmr_get_fee_rates). Returns: energy_fee_sun_per_unit, transaction_fee_sun, create_account_fee_trx, witness_pay_per_block_trx, maintenance_interval_s, next_maintenance_ms, unfreeze_delay_days, latest_energy_price.

TRON contract metadata: name, creator, whether its ABI is published on chain, the function list, code hash and whether it looks like a TRC-20. When to use: before calling a contract, or to tell an unverified contract from a published one. Do not use: for risk/backdoor analysis (that is tron_check_address_risk). Returns: found, address, name, creator, has_abi, abi_entry_count, is_trc20, functions, code_hash.

The current (tip) TRON block: number, id, parent, timestamp, transaction count and a breakdown by contract type, plus the producing witness. When to use: how busy TRON is right now, what is in the block being produced, who produced it. Do not use: to fetch an older block - a lite full node serves the tip only, historical blocks are closed. Returns: number, block_id, parent_hash, timestamp_ms, age_s, tx_count, tx_by_type, producer_address.

TRON node state from our own java-tron full node: tip height, finalized (solidity) height and how far solidity lags, peers, client version, transactions in the tip block. When to use: is the TRON node we operate synced and healthy, and how far behind is the finalized state. Do not use: for historical blocks or transactions - this is a lite full node and those APIs are closed. Returns: height, solidity_height, solidity_lag_blocks, age_s, tx_in_tip_block, peer_count, client_version.

One Tron transaction by txid: success or reverted, energy used, the call it made, and - this is the important part - any TRC-20 transfers decoded from the receipt logs (token, from, to, amount). When to use: did this USDT payment really land, for how much, from whom, and is it finalized. Do not use: for address risk scoring (tron_check_address_risk) or for blocks (tron_get_block). Boundary: our node is a lite full node with a recent window (~6 days); an empty result means 'outside our window', NOT 'the payment did not arrive' - fall back to an indexer before saying a payment is missing. Returns: found, block_number, status, contract_type, owner/to/contract_address, amount_trx, trc20_transfers[], energy_used, confirmations, finalized, retention_window.

Tron governance proposals: id, proposer, state, approvals, expiry and the parameters they would change. When to use: governance or staking questions, or checking whether a chain-parameter change is pending. Do not use: for the current parameter values (tron_get_chain_economics) or the witness list (tron_list_witnesses). Returns: total_proposals, by_state{}, top_by_approvals[] with proposal_id, proposer, approvals, state, parameters{}.

TRON super representatives and candidates: who is producing blocks, how many they produced, how many they missed, votes and websites - read from our own node. When to use: governance or staking questions about TRON, or checking the health of block production. Do not use: for account balances (tron_get_account). Returns: registered_witnesses, active_block_producers, top_by_blocks_produced[] with address, url, blocks_produced, blocks_missed, votes, in_active_set.

Read-only (constant) call against any TRON contract, executed by our own node. TRC-20 helpers balanceOf / allowance / decimals / symbol / name / totalSupply are built in; any other function can be called with a raw selector plus ABI-encoded parameter. No state is changed and no key is involved. When to use: an agent needs an on-chain value without trusting a third-party API - USDT balance of an address, token decimals, an allowance, or any view function. Do not use: to send a transaction (this never writes), or to look up a historical balance. Returns: ok, contract, function, output_type, value (decoded), value_human for 6-decimal tokens, raw, energy_used.

TRC-20 USDT settlement check by transaction id: did it land, is it irreversible, how much, from and to whom. When to use: an agent or a merchant must confirm a USDT payment before releasing value - this is the settlement check that closes an automated trade loop. Do not use: to assess the counterparty (that is tron_check_address_risk). Data source: our own java-tron full node when the transaction is inside its ~6 day retention window (sub-second, no third party); it falls back to TronGrid automatically for older transactions, and says which one answered in `source`. A local miss is never reported as 'not found' by itself. Returns: found, status (SUCCESS|FAILED), finalized (present on the solidity/irreversible chain), confirmations, blocks_until_finality, block_number, block_time_utc, amount_usdt, from, to, fee_trx, checks {}, verified and a plain-language verdict. Pass expected_to / expected_amount_usdt to have the recipient and amount matched for you. Metered: 5 free calls per day per IP, then 0.01 USDC per call over x402. There is no API key: the 402 answer carries the machine-readable terms in structuredContent.accepts[].

One Monero block header, by height or by block hash. When to use: confirming which block a transaction landed in, or inspecting reward, difficulty and weight of a block. Do not use: to list recent blocks (xmr_get_recent_blocks) or to look up a transaction (xmr_get_transaction). Returns: height, hash, prev_hash, timestamp, age_s, tx_count, reward_xmr, difficulty, block_weight, block_size_bytes, major_version / minor_version, orphan.

Monero fee estimate per kB, split into ascending priority tiers, read from our own monerod. When to use: pricing an XMR transaction, or deciding whether to pay the baseline fee to land in the next block. Do not use: for Bitcoin fees (btc_get_fee_rates) or for mempool depth (xmr_get_mempool_state). Returns: baseline_fee_per_kb_xmr plus priority_tiers[] each with {priority, fee_per_kb_atomic, fee_per_kb_xmr}. Monero has no EIP-1559 style fee market: blocks are never full, so the baseline tier normally confirms next block.

Monero mempool: pending transaction count, total bytes and fees, age of the oldest pending tx, double-spend/failing counts. When to use: is the network quiet or piling up, and roughly how long a new transaction waits. Do not use: for fees (xmr_get_fee_rates) or for a specific payment (xmr_get_transaction). Returns: pending_txs, bytes_total, fee_total_xmr, oldest_pending_age_s, double_spends, failing, tx_size_median_bytes, congestion verdict.

Monero network-wide state: network hashrate, difficulty, emission and fees over the last 720 blocks, protocol version and the next hard fork. When to use: macro questions about the Monero network - how much hashrate secures it, what is being issued, and which protocol version is active. Do not use: for our node's own health (xmr_get_node_status). Returns: height, network_hashrate_hs, difficulty, block_size/weight median and limit, emission_last_720_blocks_xmr, fees_last_720_blocks_xmr, hard_fork_version, next_fork_earliest_height, client_version.

Monero node state: height vs target, sync flag, difficulty, txpool size, database size, monerod version and peer counts. When to use: whether the XMR node is synced and healthy, or how big its chain and txpool are. Do not use: for fees call xmr_get_fee_rates; for one transaction call xmr_get_transaction; for network-wide numbers call xmr_get_network_stats. RobotBase does not index Monero wallets or ring signatures. Returns: height, target_height, synced, status, network, difficulty, txpool_txs, database_size_gb, block_size_limit, version, incoming_connections / outgoing_connections.

Monero output distribution: how many spendable outputs fall into each ring-size bucket, read from our own monerod. This is the anonymity-set side of Monero that free explorers rarely expose. When to use: reasoning about Monero privacy - how large the spendable-output set is at a given height. Do not use: to guess a 'privacy score' (we return raw counts only). Boundary: monerod's distribution data lags the chain tip by roughly 5,000 blocks, so the default range is 5,010 to 5,000 blocks back and the response states the real height and the lag. Returns: tip_height, queried_range, data_lags_tip_by_blocks, start_height, outputs_before_start, distribution[] with the raw per-bucket counts.

The most recent Monero blocks with the observed block interval. When to use: checking that XMR is producing blocks normally and at roughly the 120s target. Do not use: for one specific block (xmr_get_block). Returns: tip_height, avg_block_interval_s, target_interval_s, blocks[] {height, hash, timestamp, tx_count, reward_xmr, difficulty, weight}.

Look up one Monero transaction by txid: is it mined, in which block, how many confirmations, its size, output count and ring size. When to use: did this XMR transfer get into a block, and how deep is it now. Do not use: to learn the amount or the sender/recipient - Monero hides those by design and no third party can recover them; only the recipient (or a holder of the tx key) can prove a payment. Returns: found, confirmed, in_mempool, block_height, block_timestamp, confirmations, double_spend_seen, size_bytes, inputs, outputs, ring_size_per_input, tx_version, unlock_time, and explicitly the list of fields not disclosed by the protocol.

Historical incoming-payment ledger for one Monero address, computed with a view key. Returns per-day counts and totals, the largest common amounts, confirmations, unlocked flags and a completion flag for a 1-30 day window. When to use: a merchant or treasury needs to reconcile what actually arrived over a period, not just one txid - the thing a block explorer cannot do for Monero because amounts are hidden. Do not use: to check a single payment when you already have a tx key (xmr_verify_payment), or for balances (xmr_wallet_balance). Returns: tx_count, received_total_xmr, daily[], top_amounts[], transfers[], window.complete. The view key can read but never spend; it is used in a single-use view-only wallet and discarded.

Mining profitability calculator using live chain difficulty and the Monero tail-emission reward: gross XMR per day/month, power cost, net USD profit, break-even XMR price, break-even electricity price and hardware ROI days. When to use: decide whether a given hashrate is worth running at a power price, or to explain why XMR mining economics changed. Do not use: for pool share accounting (this service runs no pool) or price prediction. Returns: network{}, gross_xmr{}, power{}, usd{}, profitable, roi_days, break-even fields.

Verify one Monero payment cryptographically, using proof material only the recipient can provide. Give it a txid plus either the transaction key (from your own wallet: get_tx_key <txid>) or a transaction proof signature (get_tx_proof), and it returns whether that transaction really pays the given address, for how much, and how many confirmations. When to use: a merchant or agent needs to confirm an XMR payment arrived before releasing value, without trusting a block explorer - and for Monero there is no other way, because amounts and both counterparties are hidden by the protocol. Do not use: to discover payments you cannot already prove (RobotBase does not index Monero wallets, and no third party can enumerate your incoming payments), or for TRC-20 USDT settlement (that is tron_verify_usdt_transfer). Returns: verified, confirmations, received_xmr, received_atomic, in_pool. Neither the tx key nor the proof can be used to spend funds or to scan other payments.

Balance for one Monero address using a view key: total, unlocked and pending XMR, with the scanned window and a completion flag. When to use: an agent or merchant wants to know how much an address holds right now, or whether incoming funds have unlocked, without running a wallet. Do not use: for payment proof (xmr_verify_payment) or history (xmr_merchant_report). Returns: balance_xmr, unlocked_xmr, pending_xmr, scanned_to_height, chain_tip, complete. View key can read but never spend and is discarded after the call.

Balance and activity of a Bitcoin address, with no key involved. When to use: reconciling a merchant wallet, or checking whether an address received funds and how active it is. Do not use: we do not label or risk-score addresses; for a single payment use btc_get_transaction. Returns: address, confirmed/unconfirmed balance, utxo count and total, tx count, last 10 transactions. Supports P2PKH (1..), P2SH (3..), bech32 (bc1q..), bech32m (bc1p..). Ultra-active addresses may return a degraded=true response under index load rather than hanging.

Bitcoin block summary by height or hash; omit both for the chain tip. When to use: an overview of a block, or confirming which block a payment landed in. Do not use: to fetch a single payment (that is btc_get_transaction). Returns: hash, height, time, txs, size, weight, confirmations, previousblockhash.

Bitcoin fee rates for confirmation targets, plus the mempool floor. When to use: pricing a Bitcoin transaction, or deciding when a sweep will confirm cheaply. Do not use: to measure backlog depth (that is btc_get_mempool_state). There is no equivalent for Zcash: ZEC fees are protocol-fixed, see robotbase_capabilities. Returns: estimates for 1/2/3/6/12/24-block targets in BTC/kvB, recommended.fast_1_block_sat_vb / medium_3_block_sat_vb / slow_12_block_sat_vb, mempool_min_sat_vb.

Bitcoin mempool depth, capacity load and a busy/normal verdict. When to use: deciding whether now is a good moment to settle on-chain, or how large the backlog is. Do not use: to choose a fee (that is btc_get_fee_rates). Returns: txs, bytes, usage_mb, max_mempool_mb, load_pct, min_fee_btc_kvb, total_fees_btc, verdict.

Bitcoin network and node state: difficulty, network hashrate, total transactions processed and the current tx/s rate, protocol version, peer connections, relay fee, mempool size, uptime and which indexes (txindex) are synced on our node. When to use: macro questions about Bitcoin - how much hashrate secures it, how fast the chain is moving, or whether our node's indexes are complete. Do not use: for fees (btc_get_fee_rates) or mempool depth (btc_get_mempool_state). Returns: difficulty, network_hashrate_hs, chain_tx_count_total, tx_rate_per_s, client, connections, relay_fee_btc_kvb, uptime_s, indexes{}.

Look up one Bitcoin transaction by txid - the core settlement check. When to use: has this payment arrived, which block is it in, how many confirmations does it have. Do not use: for a whole block (that is btc_get_block); for an address balance (that is btc_get_address). Returns: txid, blockhash, confirmations, time, size/vsize, vin_count, up to 12 outputs with address and value, value_out_btc.

Ethereum account state from our own node: ETH balance, nonce, and whether the address is a contract. When to use: balance and account type before sizing a transfer, or telling an EOA from a contract. Do not use: for ERC-20 token balances (eth_read_contract with balanceOf) or address risk scoring. Returns: address, balance_eth, nonce, is_contract, code_size_bytes.

One Ethereum block by number, tag (latest/finalized/safe) or hash: timestamp, transaction count, gas used against the limit, base fee, block builder and size. When to use: block-level context - how full is Ethereum right now, when a block was produced, what the base fee was, or which builder produced it. Do not use: for a transaction (eth_get_transaction) or for fees across blocks (eth_get_fee_rates). Returns: found, number, hash, parent_hash, timestamp, age_s, tx_count, gas_used, gas_limit, gas_used_pct, base_fee_gwei, miner, size_bytes. Boundary: block bodies are retained from 15,500,000 (2022-09-14) only; older heights return a pruned-history error, not an empty block.

Ethereum fee market from our own Reth node: current and next base fee, priority-fee percentiles over the last blocks, and the estimated ETH cost of a plain transfer and of an ERC-20 transfer. When to use: pricing an Ethereum transaction, or deciding what tip gets included quickly. Do not use: for Bitcoin (btc_get_fee_rates), Monero (xmr_get_fee_rates) or Tron (tron_get_chain_economics). Returns: base_fee_gwei, next_base_fee_gwei, base_fee_last_blocks_gwei, priority_percentiles_gwei, estimated_cost{simple_eth_transfer_21000_gas_eth, erc20_transfer_65000_gas_eth}.

Ethereum event logs over a block range, optionally filtered by contract address and event signature, with ERC-20 Transfer logs decoded (token, from, to, amount). When to use: find transfers or contract events in a window - 'who moved this token in the last N blocks', or watch a contract's activity without an indexer. Do not use: for a single transaction (eth_get_transaction) or to scan huge ranges (max 1000 blocks per call, results truncated with a flag). Returns: filter, matched, returned, truncated, logs[] with block_number, tx_hash, address, topics, data and a decoded block for ERC-20 Transfers. Boundary: logs are read from receipts, which exist only for a rolling window (~6 days today, 1.4-8.3 days as the 50k-block static-file shard advances); older ranges are refused.

Ethereum transaction check by hash: success or reverted, gas used, fee paid, confirmations, whether the block is finalized, and any ERC-20 transfers decoded from the receipt logs (token, from, to, amount). When to use: did this ETH or ERC-20 payment actually settle, is it irreversible, who paid whom and how much. Do not use: for fee estimation (eth_get_fee_rates) or for Tron (tron_get_transaction). Returns: found, status, block_number, from, to, value_eth, gas_used, fee_eth, confirmations, finalized, erc20_transfers[]. A miss is reported as absent from *our* node only - never as proof of non-existence. Boundary: the transaction-hash index starts at block 26,170,371 (2026-10-12, only widens) while receipts cover a rolling ~6 days, so a miss on an older transaction means outside our window, not non-existent.

Read-only (eth_call) against any Ethereum contract, executed by our own Reth node. ERC-20 helpers balanceOf / allowance / decimals / symbol / name / totalSupply are built in; any other function can be called by passing its 4-byte `selector` plus ABI-encoded `parameter`. Nothing is signed or written. When to use: an agent needs an on-chain value without trusting a third-party API - a USDT/USDC balance, token decimals, an allowance, or any view function. Do not use: to send a transaction (this never writes), or for a historical balance. Returns: ok, contract, function, selector, output_type, value, value_human for known-decimals tokens.

Zcash network and node state: mempool size and bytes, node version, difficulty, network hashrate, chain progress and disk, block subsidy, and the funding streams (including Zcash Community Grants). When to use: macro questions about Zcash - how much hashrate secures it, whether our node is caught up, what the block subsidy and funding allocations currently are. Do not use: for shielded-pool supplies (zec_get_pool_status) or our node's own sync state (zec_get_pool_status / chain_status). Returns: mempool, network, difficulty, network_hashrate_hs, blockchain, height, verification_progress, size_on_disk_gb, block_reward_zec, funding_streams[], lockbox_streams[].

Zcash chain state plus the distribution of all six value pools, in one call. When to use: how much ZEC sits in each shielded pool, the shielded share of supply, and whether the ZEC node is caught up. Do not use: we do not index Zcash transactions or wallets; for node peers and disk usage call chain_status. Returns: blocks/headers/estimated_height/chain_supply/difficulty/size_on_disk_gb, node_value_pools, pools[] with {pool, balance, share_of_supply_pct, delta_1h, delta_24h}, shielded_total, shielded_pct.

The most recent Zcash blocks. When to use: checking whether ZEC is producing blocks normally and at what interval. Do not use: for shielded-pool supply (that is zec_get_pool_status). Returns: for the last N blocks (default 5, max 20): height, hash, time, difficulty.

Zcash transaction check by txid from our own Zebra node: confirmations, block height and time, whether it is in the active chain, size, and how much of it is transparent versus shielded. When to use: did this ZEC payment land, how deep is it, and does it involve shielded inputs/outputs. Do not use: to learn shielded amounts or parties - shielded values are hidden by the protocol and only a viewing key held by a party to the transaction can read them; this tool reports the shielded *shape* (spend/output/action counts) and says so explicitly rather than guessing. Returns: found, confirmations, block_height, block_hash, block_time, in_active_chain, size_bytes, transparent{inputs, outputs, total_output_zec}, shielded{sapling_spends, sapling_outputs, orchard_actions}, has_shielded_component, not_disclosed_by_protocol[].

Zcash shielded commitment-tree state: the Sapling, Orchard and Ironwood anchor roots at a given block. When to use: shielded-pool depth - how the shielded sets advance over time, or the anchor a wallet or prover needs to recognise. Do not use: for pool supplies (zec_get_pool_status) or a transaction (zec_get_transaction). Returns: block_hash, height, time, pools{sapling, orchard, ironwood}{final_root, final_state}.

Per-chain capability, retention boundary and fee model. When to use: before promising that some query is possible, or when a tool refused a query as out of range. Do not use: for live numbers (that is chain_status). Returns: one row per chain {chain, tier, name, client, state, capability}, plus the fee models we support and an honesty note. Key fact: RobotBase publishes no RPC endpoint - every read path is this MCP surface.

Live availability of every service behind the RobotBase gateway: the chain nodes, Web3 Agent Hub, AITOKENS and MCP itself. When to use: an overall platform health check, or finding which service is down. Do not use: for a single chain's node state (chain_status) or its capability boundary (robotbase_capabilities). Returns: services[] {name, ok, detail} plus an overall ok flag.

Relay an already-signed raw transaction to the network through our own full node (non-custodial: we never see a private key). When to use: an agent signed a transaction locally and wants a high-availability broadcast path into the Bitcoin network. Do not use: to build or sign a transaction (we never hold keys), or to look up a payment (that is btc_get_transaction). Returns: {accepted, txid, chain, relayed} on success, or {accepted:false, reason} - a rejected transaction is never relayed. Availability: disabled by default on this deployment; the operator enables it with RB_ENABLE_BROADCAST=1.

Live run-time state of one chain's node. When to use: is this chain's node synced, healthy or lagging; what is its height, peer count, mempool depth or client version. Do not use: fees (that is btc_get_fee_rates); a specific payment (that is btc_get_transaction); ZEC pool distribution (that is zec_get_pool_status). Returns: chain, ok, synced, height, peers, mempool, client, state; a chain that is not live yet returns served=false plus its tier instead of an error.

List every chain on the RobotBase roster with its state. When to use: the user asks which chains are supported, which nodes are online, or where to start. Do not use: for one chain's live health call chain_status; for what a node can and cannot answer call robotbase_capabilities. Returns: chains[], each {chain, tier (live|provisioning|planned), ok, synced, height, state}.

Crypto attention and market context: CoinGecko trending coins, top markets, DeFi TVL and Fear & Greed. When to use: what is trending in crypto, which tokens are moving, risk appetite, TVL leaders. Do not use: for on-chain settlement verification (use btc/eth/xmr/zec/tron tools) or price charts. Returns: {trending[], top_markets[], fear_greed, defi_tvl_top[], errors}.

Live implied probabilities from prediction markets (Polymarket primary, Manifold fallback). When to use: 'what does the market think will happen', event odds, election/crypto/macro probabilities. Do not use: for spot prices (crypto_trending) or historical search interest (attention_trend). Returns: {query, count, markets:[{question, outcomes:[{outcome, implied_probability_pct}], volume_usd, liquidity_usd, end_date, url, source}], errors}.