Ask: Graph-Native Retrieval and Cited Answers in Ontoic

Documentation Index

Fetch the complete documentation index at: https://docs.ontoic.com/llms.txt

Use this file to discover all available pages before exploring further.

When you use Ask in Ontoic, you are querying your knowledge graph — not a search engine, not the internet, and not an AI model’s training data. Every answer comes from content you have captured, and every claim in that answer links back to the specific node it came from. If your graph doesn’t contain enough information to answer a question, Ontoic tells you that and surfaces the gap as a research task rather than filling it in with a guess.

​

Zero hallucination

Ontoic’s retrieval is grounded entirely in your graph. Each answer includes numbered citations that point to the exact nodes used to construct it, and each node traces back to its original source — the PDF you uploaded, the web page you clipped, or the note you wrote. If your graph doesn’t contain an answer, Ontoic surfaces the gap as a research task rather than fabricating a plausible-sounding response. This means you can trust what Ask returns. An answer with three citations reflects three real pieces of content you captured. A “not enough information” response means your graph genuinely lacks coverage on that question — which is itself useful to know.

​

How to ask

Type any question in plain language into the Ask bar. You don’t need to use specific keywords or phrase things in a particular way — Ontoic runs the query against your graph’s semantic layer, so meaning matters more than exact wording. Results appear as a synthesised answer followed by numbered citations. Each citation links directly to the source node, so you can read the original content, see where it sits in your graph, and follow connected nodes to explore further.

​

Retrieval vs. generation

Ontoic separates two distinct steps that most AI tools collapse together:

1. Retrieval — Ontoic searches your graph for nodes that are semantically relevant to your question. This step does not involve any AI generation. The facts come from your captured sources.
2. Synthesis — Once relevant nodes are identified, Ontoic uses a large language model to compose a readable answer from those retrieved passages. The LLM arranges and summarises; it does not add information.

This distinction matters because it determines where the facts in your answer come from. In Ontoic, they always come from you.

​

Query limits

The number of Ask queries available depends on your plan: