Find what matters, not just what matches
Traditional document search works like Ctrl+F — it matches exact strings. If you search for “negligence,” you won’t find passages about “failure to exercise reasonable care,” even though they describe the same concept.
Aquiles uses semantic search. It understands what you mean, not just what you type.
How it works
Every document and reference file in your workspace is automatically processed:
- Chunked into meaningful passages (~2,000 characters, split at paragraph boundaries)
- Embedded into mathematical representations of meaning using a local machine learning model
- Stored in a vector database on your machine
When you search, your query is converted into the same mathematical representation and compared against every passage in your workspace. Results are ranked by semantic similarity — how close in meaning, not how close in spelling.
What this means in practice
...the Employee shall not engage in any activity that constitutes a violation of trust obligations owed to the Company, including but not limited to the disclosure of proprietary...
...and I understood that my role required me to act in the best interest of the organization at all times, which is why I consulted legal before making that decision...
...this conduct represents a clear failure to uphold the duties of loyalty and care expected of a senior officer in this capacity...
- Search for “breach of contract” and find passages discussing “failure to perform contractual obligations”
- Search for “proximate cause” and surface discussions of “foreseeable consequences”
- Search for “damages” and retrieve passages about “financial harm,” “quantifiable losses,” or “compensatory relief”
- Find evidence you didn’t know existed in your own files
Privacy by design
All embedding generation happens on your machine using a lightweight local model. No document content is sent to any external service for indexing. The vector database is stored locally alongside your case data.
Semantic search works entirely offline. Your case files never leave your device for search to function.
Powering everything
Semantic search isn’t just a feature you use directly — it’s the engine behind Aquiles’ AI. When you ask a question in the AI chat, the system uses semantic search to find the most relevant passages across all your documents, then feeds those passages to the AI for synthesis.
This is why AI answers in Aquiles are grounded in your evidence rather than generated from generic training data. The search finds the evidence. The AI explains what it means.