← Back to Why was a vocabulary read refused?
How do I read a term by name, and the JSON-LD context?
Short answer
Read a type at https://veda.gilb.com/base/t/<Name>, a field at /base/a/<name>, and the JSON-LD context at /base/context.jsonld, with no sign-in.
Before you start
The term's exact name, capitals included.
Words used here (3)
- field
- A field is one named fact on a record, such as its title or a due date; its name comes from GilbVeda.
- record
- A record is one piece of data with its own address, such as a person, a place, a thing or a note.
- short name
- A short name is the one word you give one AI on one computer, such as gilb-example; it keeps that AI's key and tools apart.
Steps
Read a type by its name.
Fetch the type's document from GilbVeda, the shared list of terms, with no token:
GET https://veda.gilb.com/base/t/<Name>More about this
A term is the shared name GilbVeda keeps for a kind of thing, a detail of it, a link between two things, or a species. A type is a kind of thing, such as Note. The answer is JSON-LD, JSON with a fixed set of names, of type application/ld+json. It holds the term's facts, such as a:label, its name; a:does, its meaning; a:ontologyStatus, Suggested, Approved or Deprecated; a:module; and a:subTypeOf, the term it is a kind of. Every status is served. The name is matched exactly, capitals included: /base/t/note answers 404 where /base/t/Note answers 200. A name GilbVeda does not have answers 404 NotFound. Ask veda.gilb.com: the same path on https://id.gilb.com answers 404.
Read a field's term by its name.
Fetch the term that defines a field, with no token, using the field's name without its a: prefix:
GET https://veda.gilb.com/base/a/<name>More about this
A field is one named fact on a record, such as its title or a due date; its name comes from GilbVeda. A record is one piece of data with its own address, such as a person, a place, a thing or a note. So a:title is read at https://veda.gilb.com/base/a/title. The answer has the same shape as a type's and adds a:encoding, how the field's value is written, such as inline-literal:text, and a:fieldOf, the types the field belongs to. Its a:typeUri is the same term's address as a type, here https://veda.gilb.com/base/t/Title. Capitals count here too: /base/a/Title answers 404 NotFound.
To read several types at once, name each one with type=.
Name each type as type=, at most 100 different names, in one request:
GET https://veda.gilb.com/base/vocab?type=<Name>&type=<Name>More about this
The answer is one JSON-LD document. The terms found are in @graph, each with its status, and a name GilbVeda does not have is listed under absent. Write the bare name: type=t:Note answers 400 type_token_unresolved and, where it can, names the name you probably meant. More than 100 different names answer 400 type_filter_too_large, and nothing is cut. type= does not go with limit, cursor or nodeKind: together they answer 400 SelectorCombinationUnsupported.
Read the JSON-LD context.
Fetch the context of the vault's host whose documents you read, with no token, such as GilbVeda's:
GET https://veda.gilb.com/base/context.jsonldMore about this
A vault is the place where a person or a group keeps their records, each at its own web address. The context is the part of a JSON-LD document that says what its short names stand for. Here @vocab is https://veda.gilb.com/base/a/, so a bare name such as title stands for https://veda.gilb.com/base/a/title; t: stands for https://veda.gilb.com/base/t/ and a: for https://veda.gilb.com/base/a/. Each vault's host serves its own, built from its own address: at https://id.gilb.com/base/context.jsonld, @vocab is https://id.gilb.com/base/a/. A record read at its /base/e/ address names its host's context in @context; a term's document carries the context written out; a read through /lws names https://www.w3.org/ns/lws/v1 instead. A host the platform does not serve answers 404.
Deeper
None of these reads needs an account, a token or a proof, and none changes anything. A web page on any site may call them: they answer access-control-allow-origin *. A term, a list and the context may be kept for an hour (cache-control public, max-age=3600). A term's answer carries an etag; send it back as If-None-Match, and a term that has not changed answers 304 with no body.
The published vocabulary can also be read as one module, at https://veda.gilb.com/base/vocab/<module>. Today the only module served is planguage-core, and it is the whole published vocabulary, its Approved terms, in one document: 175,313,562 bytes on 2026-11-11. Any other module answers 404, base-internal, the module of terms GilbPlatform keeps for its own use, included. To read many terms, name them with type= or page the list; the answer on listing, browsing and searching the vocabulary shows how.
When two terms claim the same address, a read of a type, of several types with type=, or of the module answers 409 IdentityCollision and names the address, instead of choosing one. A term's document may hold more facts than the ones named here, such as a:synonym, its other names, and a:example.
To find a term's exact name, search the vocabulary; the answer on listing, browsing and searching the vocabulary shows how. These reads change nothing: Agent.Concept, the agent that finds and makes GilbVeda's terms, writes them, and the answer on minting a term shows how. Species are kept apart, in GilbLife, GilbPlatform's list of species.
The full reference is the Developer Guide, GilbPlatform's guide for people who build apps on it, in its chapter Dev.Guide.Vocabulary, https://appbuilder.gilb.com/base/e/06gkzpwtj1wm58yf6zng0xdtbm. It can be read with no sign-in, as data for programs and AIs rather than as a web page.
Source: GilbPlatform 498d29a0 (live at https://id.gilb.com/version, read 2026-11-11): src/routes/vocab.ts:1-13 (route list), :55-75 (public by design; base-internal never served), :79-81 (cache-control public, max-age=3600, ETag), :364-376 (serveJsonLd: etag, If-None-Match 304, application/ld+json), :183-191 (collisionReply: 409 IdentityCollision), :771-806 (GET /base/t/:Type: host-routed IRI, exact typeUri lookup, 404 notFoundBody, any status served, Cycle145 note), :823-855 (GET /base/a/:slug: resolved through a:predicateAttribute bindings, subject = the attribute IRI), :931-940 and :1100-1123 (?type= batched read on the canonical door, any status, absent declared), :979-985 (SelectorCombinationUnsupported), :1196-1225 (GET /base/vocab/:module: base-internal 404, empty module 404); src/vocab/term-reader.ts:1289-1308 (listPublishedTerms: a module other than planguage-core returns [] -> 404; Approved only); src/read/type-name-selector.ts:39 (MAX_TYPES 100), :71-110 (type_token_unresolved with probablyMeant, type_filter_too_large, never truncated); src/auth/errors.ts:86-90, :151-155 (error names); src/server/app.ts:244-256 (4xx body error, message and details); src/routes/_http.ts:26-28, :52-59 (404 NotFound No resource at ...; unregistered Host 404); src/routes/read.ts:739-764 (GET /base/context.jsonld: public, host-routed, buildContext(routedHost), cache-control public, max-age=3600, no etag); src/jsonld/context.ts:160-175 (@vocab https://<host>/base/a/, t:, a:, base:, role:), :244-246 (contextUrl), src/jsonld/frame.ts:388, :441 (a /base/e read names contextUrl(host) in @context); src/lws/jsonld.ts:39 (LWS_CONTEXT_URL https://www.w3.org/ns/lws/v1). the code at 498d29a0 shows: no match in src/routes/vocab.ts (no auth step on these reads). Public reads, no token, 2026-11-11: /base/t/Note 200 application/ld+json, cache-control public, max-age=3600, etag, access-control-allow-origin *; /base/a/title 200 (a:encoding inline-literal:text, a:fieldOf, a:typeUri https://veda.gilb.com/base/t/Title, a:predicateAttribute a:title); /base/context.jsonld 200 (@vocab https://veda.gilb.com/base/a/, no etag); /base/t/NoSuchTypeZz 404 {error:NotFound,message:No resource at ...}; /base/vocab?type=Note&type=Title&type=NoSuchTypeZz 200, complete true, absent [NoSuchTypeZz]; ?type=t:Note 400 type_token_unresolved, probablyMeant Note; ?type=Note&limit=3 400 SelectorCombinationUnsupported; /base/vocab/skill-esco-thin 404, :26Z /base/vocab/base-internal 404; HEAD /base/vocab/planguage-core 200, content-length 175313562; /base/t/note 404, /base/a/Title 404; https://id.gilb.com/base/t/Note 404; https://id.gilb.com/base/context.jsonld 200, @vocab https://id.gilb.com/base/a/. Dev.Guide.Vocabulary (A 06gkzpwtj1wm58yf6zng0xdtbm, pack GilbPlatformDeveloperGuide, 2026-11-11; read for agreement, not copied) Names, Find, Encodings, Lists; Dev.Guide.Overview (A 06gkzpwebxv8ncqr4qd7egmt58, 2026-11-11) WhatItIs (JSON-LD (JSON with a fixed set of names)). Also: Agent.Concept.
For developers
curl
read-terms.sh
# Read one type and one field's term. No token is needed.
curl -s "https://veda.gilb.com/base/t/<Name>"
curl -s "https://veda.gilb.com/base/a/<name>"
# Several types in one request, at most 100 names.
curl -s "https://veda.gilb.com/base/vocab?type=<Name>&type=<Name>"
# The JSON-LD context of GilbVeda's host.
curl -s "https://veda.gilb.com/base/context.jsonld"
TypeScript
read-term.ts
const VEDA = "https://veda.gilb.com";
export interface TermSummary {
address: string;
name: string;
meaning?: string;
status?: string;
encoding?: string;
}
// Reads a type ("t") or a field's term ("a") by its exact name, capitals included.
// Answers null when GilbVeda has no term of that name. No token is needed.
export async function readTerm(kind: "t" | "a", name: string): Promise<TermSummary | null> {
const res = await fetch(`${VEDA}/base/${kind}/${encodeURIComponent(name)}`);
if (res.status === 404) return null;
if (!res.ok) throw new Error(`term read answered ${res.status}`);
const doc = (await res.json()) as Record<string, unknown>;
const text = (key: string): string | undefined =>
typeof doc[key] === "string" ? (doc[key] as string) : undefined;
return {
address: String(doc["@id"]),
name: text("a:label") ?? name,
meaning: text("a:does"),
status: text("a:ontologyStatus"),
encoding: text("a:encoding"),
};
}
Checked against GilbApp d529620.