# dictionary A RocketRide text-processing node that uses an LLM to extract a glossary of domain-specific terms from documents. ## What it does Reads text and uses a connected LLM to extract a dictionary of domain-specific terms, acronyms, and company-specific vocabulary, including terms that are ambiguous or whose company meaning differs from common usage. Each extracted term is emitted as a separate document whose content is a `{"term": "...", "description": "..."}` JSON object. Both plain text and table content are handled the same way: each incoming chunk is sent to the LLM with a structured prompt (`expectJson: true`), the LLM returns a single JSON array of definitions, and the node writes one output document per definition. Output documents are stamped with an incrementing `chunkId` (reset for each input object) and marked as non-table content (`isTable: false`, `tableId: 0`). The node has no Python package requirements of its own: it relies entirely on the separately installed AI module. --- ## Connections | Connection | Required | Description | |------------|-------------|--------------------------------------| | `llm` | yes (min 1) | LLM used to extract and define terms | --- ## Configuration ### Lanes | Lane in | Lane out | Description | |---------|-------------|--------------------------------------------------------------| | `text` | `documents` | Extract term/description pairs, emit one document per term | No configuration fields. Connect an LLM and wire text in. --- ## Usage Place this node after a text source (for example, a parse or preprocessor node) and before a vector store. The output documents can be ingested into a store and queried later to enrich LLM context with domain-specific definitions. The LLM prompt instructs the model to: - Pull out company-specific terms, acronyms, and information that may help answer user questions. - Highlight any terms whose meaning in the document differs from common usage. - Merge definitions from multiple documents into a single JSON array before returning. - Return definitions that may overlap when a term has more than one relevant sense. Example output document for a single extracted term: ```json {"term": "Red loan", "description": "Credit score less than 650 and delinquent by 60 days or more"} ``` --- ## Schema _No configuration fields._ ## Source [ View source](https://github.com/rocketride-org/rocketride-server/tree/develop/nodes/src/nodes/dictionary)