Try the data problem for yourself.

The claim: claims grounded in your records link to the supporting email or message, while inference and missing evidence remain visible instead of being presented as sourced fact.

Verify it: ask something only your mail knows ("what did the plumber quote me?"), click the citation, and read the original message. Then ask something outside the record and inspect how the answer labels its evidence limits.

The claim: you can choose another supported AI while keeping your Talavine data and conversation. The available models differ in capability, cost, and privacy settings; answers and citations can differ too.

Verify it: ask a question, choose another configured model, and ask again. Check that your saved records remain available and compare the answers against those records. For local AI, review the settings for the whole workflow; changing the chat model alone does not disable every cloud connection.

The claim: a useful answer may need a text message as well as an email or document. Talavine brings supported sources together, subject to account access, device permissions, and workplace policies.

Verify it: on a Mac, connect Apple Messages with the required permissions and ask a question that spans a text conversation and an email thread. Open the supporting records and check whether anything is missing.

The claim: for repeated jobs, the AI authors a small, typed, sandboxed C# tool once — then the job runs as deterministic code. A portfolio tracker's refresh dropped from ~42 model calls to zero.

Verify it: in Vibes, describe a stock or weather tracker, let it build, then open the app's “Inside” view: read the generated tool's code, its declared parameters, and its last-run record — and refresh the rows without a single model call.

The claim: work pinned local-only can never fall back to a cloud model (it stops and says so instead); unproven models can't silently take over the mail pipeline (calibration-gated routing); paying per message is off by default; and the docs state plainly that Claude Code subscription routes do not count as "local."

Verify it: read the white paper and routing page, then check the behavior in the app — set a function to local-only with no local model installed and watch it stop instead of silently going to the cloud.

The claim: give a helper a clear job, the information it needs, and limits on what it may do. It can prepare work, gather updates, and report what needs your attention. Desktop work requires an awake, connected computer.

Verify it: use a test profile with client emails and meeting notes. Ask a helper to draft a follow-up without sending it. Review the draft, compare it with the records, and check the activity log and pause controls. The client-work example shows a starting request.

On the record: “I started building Talavine because I found the agent gateways everyone was excited about unusable. This is the version I could trust on my own machines.” — Scott Crossen, creator of Talavine. Quote freely.