other

Takes chain, token pair, input amount and slippage tolerance, finds the pool the trade would most likely use and reads its live state. The maximum the caller can lose is arithmetic and exact: the gap between the quoted output and the minimum output their own tolerance permits. The attacker profit is an estimate and labelled as one — the profit-maximising front-run size is searched for inside a constant-liquidity pool model, bounded by the largest front-run the caller own tolerance still permits,

Resolves a pool in a public yield directory by chain, protocol and pool id or symbol, then rebuilds the number a buyer actually needs. Advertised total APY and a conservative APY are reported separately: the conservative figure uses the trailing seven-day realised base yield and deliberately excludes token emissions, because emissions are only worth what the reward token is worth. Every reward token is priced at the moment of the call and that price is stated as the assumption it is; the end of

Converts to FLAC, WAV at 16 or 24 bit, MP3, Opus or AAC, with sample rate, channel count and quality all settable and all reported back. What separates this from a bare transcode is the verdict on what the conversion actually did. The spectral cut-off of the source is measured first, so a source that already passed through a lossy stage is recognised before it is wrapped in a lossless container: converting an MP3 to FLAC multiplies the file size by roughly ten and restores nothing, and the…

Classifies only three broad German-language varieties: de-DE Germany German, de-AT Austrian German and de-CH Swiss German. The model works on acoustic speech embeddings and does not read a transcript. It returns probabilities for all three classes, the strongest candidate, audio quality warnings and unknown whenever confidence or recording quality is insufficient. The result is a statistical variety estimate, not a statement about a speaker's origin, identity, nationality or place of residence.

Walks the visible book of every requested exchange for the given size, in both directions, and works out what remains of the raw spread: gross_spread_bps from top of book, then the taker fee of the entry tier on both sides, then the slippage the size causes on the buy book and on the sell book. net_spread_bps is the headline number, the gross value stands next to it because on its own it is misleading. Transfer costs are the part no public exchange interface publishes without a key: withdraw fee

Returns the sections of an official drug label sorted by clinical significance rather than document order, with the text passed through word for word. Nothing is summarised and nothing is shortened: a summary is by definition an omission, and an omitted contraindication is the kind of omission that matters. A caller who needs a shorter version can produce one from the complete text supplied here, which is the right order of operations.

Checks vouchers of a payment channel by computation only - nothing is ever claimed, redeemed or refunded. Signatures are verified cryptographically against the address each voucher claims to be signed by, in EIP-712 form when the typed data is supplied, otherwise as a plain message signature. Double spending is the expensive case and is checked in full: every voucher id occurring more than once is reported with all its occurrences and whether the duplicates agree on the amount. Sums are added up

Will it rain in San Jose during your event window — Event-window weather odds for San Jose: pass a start and end time (or omit them for the next six hours) and get rain probability and expected conditions across the window — built for scheduling decisions. Same shape as /v1/forecast/window with the San Jose coordinates built in.

Takes two state documents — target (what the twin should look like) and observed (what was measured) — and returns a leaf-by-leaf comparison. Every leaf path carries one of five states: in_tolerance, out_of_tolerance, missing_in_observed, unexpected_in_observed or type_mismatch.

Turns opaque call data into something an agent can reason about before signing. The method signature is resolved from the verified source of the contract first, then from this hub's own list of common signatures, then from the open selector directory — and the answer always names which of the three was used, because a four-byte selector is short enough to collide and colliding entries are listed. An unknown selector stays unknown: parameters are returned raw and nothing is guessed. Approvals and

Accepts three forms of recipient. A bech32 lnurl is decoded to its callback URL, a Lightning address is resolved through the well-known lnurlp path, and a Nostr public key is resolved by reading its kind 0 profile from relays and taking the lud16 or lud06 field it publishes.

Plans the division of a long input for a given model. It reserves room for the system prompt, the expected answer and a safety margin, then works out how many chunks are needed and where to cut, preferring paragraph boundaries over line and sentence boundaries and never cutting inside a word. Optional overlap carries context across the seams.

Liest Format, Masse, Farbraum, Bittiefe, Alphakanal, Bewegtbild, ICC-Profil und Metadaten direkt aus dem Dateikopf von PNG, JPEG, GIF, WebP und SVG. Prueft gegen selbst gesetzte Anforderungen an Groesse, Seitenverhaeltnis, Format und Metadatenfreiheit und meldet Fallstricke wie CMYK, EXIF mit Ortsangabe oder sehr grosse Pixelmengen.

Reads the public perpetual data of Binance USDⓈ-M Futures, Bybit linear and OKX swaps for one BASE/QUOTE pair. Settlement intervals differ per exchange and per contract, so annualized_pct is only computed after normalising to the interval that exchange actually uses, and the formula is part of the answer. current_rate and predicted_rate are strictly separate: where an exchange publishes no forecast, predicted_rate stays null and the reason is named instead of reusing the current rate. next_settl

Converts each OpenAPI operation into an MCP tool. Path, query and header parameters and the JSON request body are merged into one input schema, because MCP tools take exactly one; a name collision between a body field and a parameter is reported rather than silently overwritten. Internal $refs are expanded inline so each tool schema stands alone — an MCP client cannot resolve them.

Turns SenML (JSON or CBOR) and LwM2M TLV into one record shape: resolved name, unit, absolute timestamp, typed value. Optionally puts every series on a shared time axis with a chosen interval and fill rule (last value, bucket mean, linear). Works only on the payload you pass: no device is contacted, nothing is stored, the same request returns the same bytes. TLV carries no type on the wire; unmapped resource ids come back as opaque base64 with a warning instead of a guess.

Decodes a JWT and checks it in two layers. Structure and claims: expiry, not-before and issued-at against the current time with the clock skew you allow, issuer and audience against the values you expect, and the token lifetime. Signature: HMAC with a shared secret, or RSA, RSA-PSS and ECDSA against a PEM key or a JWKS you point at, with the kid matched to the right key.

Providers each build their signing string differently — some sign only the body, others prepend a timestamp, an id, or a version marker in a fixed order — and getting that order wrong produces a mismatch that looks exactly like an attack. This checks the signature the way the named provider actually specifies, with constant-time comparison, and returns the string that was signed so a mismatch can be debugged instead of guessed at.

Reads an OpenAPI 3.x document and reports what would stop a machine consumer. Errors: missing openapi, info or paths; a path that declares no method; an internal $ref that resolves to nothing; a parameter without a name, location or schema; a request body without content.

Builds the cohort table from the events you pass: users grouped by the period of their first event, counted again in each following period. Returns the table per cohort, a retention curve across all cohorts, and, where events carry a value, cumulative value and value per user. The curve counts only cohorts old enough for a given period, so a young cohort is not scored as churn. Day, week and month are real calendar periods. Nothing is fetched, nothing is stored.

Opens a TLS connection and reports the chain the server presents. Per certificate: subject and issuer, validity window with days remaining, serial, SHA-256 fingerprint, key type and size, and the subject alternative names. Above that it answers the questions a monitor asks: is the requested hostname covered by the leaf certificate including wildcard rules, how many days until the nearest expiry, is any certificate self-signed or signed with a weak algorithm, does the server send its…

Answers the timing question an agent has before it schedules anything against a regulated exchange: is the main session running, and if not, when does it start. Covers XETR, XFRA, XNYS, XNAS, XLON, XPAR, XAMS, XSWX, XTKS and XHKG with pre trading, main session, post trading, half days and the holidays of 2026 and 2027. Time zones are resolved through IANA names and Intl at the queried instant, never through a fixed offset to UTC, because summer time starts and ends on different dates in the Euro

Looks up the declared licence of each package through deps.dev and sorts the result by the only question that matters for a closed, sold product: what does this licence demand. Permissive licences need attribution. Weak copyleft affects changes to the library itself. Strong copyleft can force disclosure of the whole work on distribution. Network copyleft such as AGPL and SSPL bites on operating the software as a service, which is the sharpest case for a paid API.

Two checks in one. First the arguments are validated against the tool input schema and every violation is reported with its path and a sentence saying what to change. Second, values that satisfy the schema but look wrong are flagged: an unset required field filled with a placeholder like "string" or "TODO", a number at a suspicious order of magnitude, an address that fails its checksum shape, a destructive flag set to true, an empty string where content was expected.