|
| 1 | +--- |
| 2 | +title: Experimental Features |
| 3 | +description: A home for in-progress experiments in Wails v3 — what they are, why they exist, and how to give feedback. |
| 4 | +sidebar: |
| 5 | + label: Overview |
| 6 | + order: 0 |
| 7 | + badge: |
| 8 | + text: Experimental |
| 9 | + variant: caution |
| 10 | +--- |
| 11 | + |
| 12 | +import { Aside, CardGrid, LinkCard } from '@astrojs/starlight/components'; |
| 13 | + |
| 14 | +<Aside type="caution" title="Here be experiments"> |
| 15 | +Everything in this section is, by definition, an experiment. Features here are |
| 16 | +opt-in, off by default, and may change shape, get renamed, or be removed |
| 17 | +entirely between releases. Don't build anything load-bearing on them without |
| 18 | +being ready for churn. |
| 19 | +</Aside> |
| 20 | + |
| 21 | +## What "Experimental" means |
| 22 | + |
| 23 | +Wails ships experiments in the open. An experiment is an idea we think is |
| 24 | +promising enough to put in your hands early — but one we haven't fully |
| 25 | +committed to yet. We're publishing it *because* we want to learn from real use |
| 26 | +before deciding whether it becomes a permanent, supported part of Wails. |
| 27 | + |
| 28 | +That means a few things are true of everything in this section: |
| 29 | + |
| 30 | +- **It's opt-in.** Experiments never change the default behaviour of `wails3`. |
| 31 | + You turn them on deliberately (usually with an environment variable or a |
| 32 | + build flag), and with them off, nothing about your existing workflow |
| 33 | + changes. |
| 34 | +- **It might not survive.** Some experiments graduate into stable features. |
| 35 | + Others get reworked beyond recognition, or dropped. We'd rather try things |
| 36 | + in public and learn fast than ship only what we're already certain about. |
| 37 | +- **The API isn't frozen.** Names, flags, defaults and behaviour can shift |
| 38 | + between releases while an experiment is finding its shape. Release notes will |
| 39 | + call out the changes, but don't expect the stability guarantees that stable |
| 40 | + features carry. |
| 41 | + |
| 42 | +## We want your feedback |
| 43 | + |
| 44 | +This is the important part. Experiments live or die by what we hear back from |
| 45 | +people actually using them. If you try one of these, we genuinely want to know: |
| 46 | + |
| 47 | +- Did it work for your project? Where did it fall down? |
| 48 | +- Was it faster / clearer / nicer — or not worth the switch? |
| 49 | +- What would have to be true for you to use it by default? |
| 50 | + |
| 51 | +The most useful feedback is concrete: what you ran, what you expected, what |
| 52 | +actually happened. Each experiment has its own thread in the **Experiments** |
| 53 | +category on GitHub Discussions: |
| 54 | + |
| 55 | +<CardGrid> |
| 56 | + <LinkCard |
| 57 | + title="Experiments Discussions" |
| 58 | + description="Find the thread for the experiment you're using and tell us how it worked out, what broke, or what's missing." |
| 59 | + href="https://github.com/wailsapp/wails/discussions/categories/experiments" |
| 60 | + /> |
| 61 | +</CardGrid> |
| 62 | + |
| 63 | +## Current experiments |
| 64 | + |
| 65 | +<CardGrid> |
| 66 | + <LinkCard |
| 67 | + title="Wake" |
| 68 | + description="An alternative, Wails-aware build runner for your existing Taskfiles. Faster incremental builds, structured output, parallel by default." |
| 69 | + href="/experimental/wake" |
| 70 | + /> |
| 71 | + <LinkCard |
| 72 | + title="LLM Control (MCP)" |
| 73 | + description="A built-in Model Context Protocol server that lets LLM agents inspect, test, and drive a running Wails app — no user code required, enabled with a build tag." |
| 74 | + href="/guides/mcp-service" |
| 75 | + /> |
| 76 | +</CardGrid> |
0 commit comments