VERA is the first agent in the ThreeClaws trio, and the one most people meet first. Her job description is simple to state and surprisingly deep in practice: read things carefully, remember what matters, and draft what comes next.
What VERA is for
Think of the hours in a week you spend gathering context before you can actually do work. Reading an inbox. Skimming a case file. Pulling together notes from a long-running client relationship so you can write one three-paragraph email. VERA is built for that layer.
A typical day with VERA looks like:
- Morning brief. She reads overnight email, pulls out the items that need a human, and drafts replies to the ones that don’t.
- Pre-meeting prep. Given a client name, she assembles everything she knows — notes, past drafts, open questions — into a short briefing.
- Document review. Drop a 40-page contract, a research paper, or an intake form on her, and get back a structured summary with the parts that need your attention flagged.
- Draft generation. Memos, proposals, follow-ups, first-pass research notes. She writes the boring 80%, you write the 20% that requires judgment.
How she’s different from a chatbot
The difference between VERA and a general-purpose cloud chatbot is memory and locality.
A chatbot forgets you every conversation. VERA remembers every document she has ever read for you, because those documents never left your office.
That’s the key. She builds a private, local knowledge base out of the actual files on your machine. When you ask her a question, she’s not guessing from training data — she’s answering from your own history.
Working with her teammates
VERA doesn’t operate alone. She hands structured work over to Nexus, the workflow agent, when something needs to become a task, a calendar event, or a tracked case. She hands drafts to Clara, the client communication agent, when something needs to go out the door in your voice.
The three of them form a loop:
- VERA reads and drafts.
- Nexus routes and schedules.
- Clara polishes and sends.
Each one does its own job well, and each one runs on the same local hardware.
Getting started
If you’re evaluating whether a local AI agent fits your practice, VERA is the easiest place to start. She doesn’t touch outgoing communication, she doesn’t make scheduling changes on her own, and she’s easy to sandbox to a single folder while you get a feel for her.
We’ll go deeper on Nexus and Clara in the next two posts in this series. If you want to be notified when those go up, there’s a contact form on the homepage — no mailing list, just a note from us when they land.