Software you can actually use.

Write in plain markdown, then let the vault do the remembering: search by meaning rather than keyword, get flashcards made from your own notes, and see a map of what you actually write about. The AI runs on your device, so nothing you write leaves it.

For anyone who takes notes and never reads them again.

Point it at a GitHub repository and hand it a task. Planner, engineering and full-stack agents split the work, run it in parallel, and come back with a pull request you can review like any other.

For developers with a backlog of small, well-defined issues.

Drafts the proposal, then watches the project against it — flagging scope creep as it happens and keeping a tamper-evident record of what was actually agreed, so the awkward conversation has evidence behind it.

For freelancers and agencies who have lost money to scope creep.

Photograph a plant and get an identification with care advice tuned to your local weather. List what you have spare, find someone near you who wants it, and arrange the handoff in chat.

For anyone who has propagated one too many pothos.

Connects to a GitHub repository and scores each pull request two ways: the sentiment of the review conversation, and the estimated carbon impact of what the change actually does.

For teams who want their review culture and footprint measured.

Analyses patterns in what people choose and lets you interrogate the result in conversation. Falls back across models automatically, so it answers even when a provider is down.

For anyone exploring preference data without writing queries.

A dashboard study: realtime visualisation, dynamic content, and a lot of attention paid to micro-interactions and accessible UI patterns.

For a look at the front-end craft rather than a tool to adopt.