webly — Claude Code and Codex skill: what it does, how to install
webly publishes a site or a folder of files at a live URL without an account: the anonymous copy lasts 24 hours, then you claim it and edit over MCP.
Apache-2.0 ★ 1,596 updated 2026-10-29
npx skills add hashgraph-online/awesome-codex-plugins --skill webly -a claude-codeWhat webly does
Ask the agent to publish and it puts a static site or a set of files online at a working address. No sign-up is needed for that first pass. The anonymous copy stays up for 24 hours, and claiming it into an account is possible within 7 days.
Every deterministic step runs through scripts/webly.mjs, a helper with no dependencies that wants Node 20+. The webly doctor check only reads: it reports the saved token, site status and links, whether the MCP server https://api.webly.ai/v1/mcp is configured, and which step is still open. On older Node the helper will not start.
The API contract lives in llms.txt, and when it disagrees with the live API, the live API wins. File sharing covers up to 500 MiB across links, each good for 24 hours. Sign in over OAuth through MCP and the site becomes permanent, with edits through MCP tools. Drafts wait for your approval; rollback, custom domains, forms and analytics come with it.
What it helps with
- Put a finished site online and get a link back
- Send a folder of files as a single download link
- Claim an anonymous site before the window closes
- Live edits over MCP with draft approval
- Point your own domain at a Webly site
Example prompts for the agent
- Publish this site and give me the link
- Share the
distfolder as a download link - Claim the site into my Webly account and connect my domain
How to install webly
1Skills CLI (Vercel)
npx skills add hashgraph-online/awesome-codex-plugins --skill webly -a claude-code2Manually: put the files in place
~/.claude/skills/webly/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/webly/
1Skills CLI (Vercel)
npx skills add hashgraph-online/awesome-codex-plugins --skill webly -a codex2Manually: put the files in place
.agents/skills/webly/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add hashgraph-online/awesome-codex-plugins --skill webly -a cursor2Manually: put the files in place
.cursor/skills/webly/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install hashgraph-online/awesome-codex-plugins plugins/Webly-AI/webly-plugin/skills/webly2Manually: put the files in place
.github/skills/webly/SKILL.md1Skills CLI (Vercel)
npx skills add hashgraph-online/awesome-codex-plugins --skill webly -a gemini-cli2Platform CLI
gemini skills install https://github.com/hashgraph-online/awesome-codex-plugins --path plugins/Webly-AI/webly-plugin/skills/weblyManually: put the files in place
.devin/skills/webly/SKILL.mdThe legacy .windsurf/skills/ path still works
Claude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about webly
How do I install webly in Claude Code?
Run npx skills add hashgraph-online/awesome-codex-plugins --skill webly -a claude-code; the skill file then sits at ~/.claude/skills/webly/SKILL.md.
What does webly do?
It publishes an agent-built site or a set of files at a live URL with no account. The anonymous copy runs 24 hours; after OAuth sign-in the site moves into the account and is edited through MCP tools.
How long does an unclaimed site stay online?
24 hours. You can claim it within 7 days, and from then on it is permanent, with drafts, rollback, forms and analytics.
Source text
SKILL.md · Apache-2.0
---
name: webly
description: >-
Webly hosts what agents make: publish a website, or share files and folders,
at a live URL in seconds with no account, then let the person sign in and claim it
to keep it, edit it over MCP, and publish drafts they approve. Use when asked
to "publish this", "host this", "deploy this", "put this online", "make a
website", "share this as a web page", "give me a link to this site", "share
these files", "send these files", "give me a download link", "keep my
site", "claim my site", "make a Webly account", "connect Webly", "connect
my domain", "use my own domain", or when
working on a site hosted on Webly (*.webly.site). Also use at the start of any
session where a Webly site or an unfinished Webly step is waiting on this
machine.
license: Apache-2.0
metadata:
publisher: Webly
homepage: https://webly.ai
docs: https://api.webly.ai/docs
api-contract: https://api.webly.ai/llms.txt
mcp-server: https://api.webly.ai/v1/mcp
---
# Webly
Webly hosts the websites and shared files and folders that agents publish. Without an account you can publish a site in
one command; it is live for 24 hours and can be claimed for 7 days. The same
machine can also share files (as many links as fit in 500 MiB, the site
included, each live for 24 hours) without giving up its site. Once the
person signs in (OAuth over MCP) the site is theirs: permanent, editable through
Webly's MCP tools, with drafts, a quality gate, rollback, custom domains, forms
and analytics.
`{api}/llms.txt` is the full contract, where `{api}` is the `api` that
`webly.mjs doctor` reports (`https://api.webly.ai` unless `WEBLY_API_URL` is
set). Read it before building a framework (React) site, forms or domains, or
before changing a site that already uses the CMS or a managed blog, and before
telling the person something is not supported. If this file and the live API disagree, trust the API.
The helper that does every deterministic step is `scripts/webly.mjs` in this
skill's folder (zero dependencies, Node 20+). Below, `webly` means
`node <this skill's folder>/scripts/webly.mjs`.
## Language
Reply in the person's language. Keep URLs, commands, tool names and error
reasons exactly as returned.
## Step 0: find out where we are (every time, first)
A new session, `/clear` or a restart wipes the conversation, so all state lives
on disk and at the API. Rebuild it before doing anything:
1. Run `webly doctor`. It is read-only and never creates a token. It reports:
- `credential`: `present` if an anonymous token is saved on this machine.
- `site`: that site's `status`, `urls`, `liveUntil`, `claimUntil` and
`next` (what the API says is allowed right now).
- `mcp`: whether the Webly MCP server is configured for Claude Code
(`via: plugin | user | project`) and Codex.
- `project`: the site(s) the project in the current folder is linked to
(`.webly/project.json`, see **Linked folders**), or `error` if that file
can't be read.
- `signedIn`: `true` if `webly login` saved a sign-in for this API.
- `pending`: a step an earlier session asked you to finish, e.g. `claim`.
- `updatedFrom`: set on the first run after this skill was updated. Tell
the person in one line which version they are now on.
- `update`: set when a newer version of this skill is published. Run its
`command` (it matches how this copy was installed), tell the person in one
line, then carry on with the task using this copy. The new one loads next
session.
2. Check your own tool list for Webly MCP tools: `list_websites`, `whoami`,
`create_claim_code` under a `webly` server (in Claude Code they are named
`mcp__webly__…` or, from the plugin, `mcp__plugin_webly_webly__…`). A server
that is configured but not in your tool list is **not loaded**.
Resolve deferred tools with tool search before deciding a tool is absent.
Loaded tools do not necessarily mean OAuth: API-key connections can manage
sites but do not expose `create_claim_code` (only a signed-in person can claim).
3. **The folder's link comes first.** If `project.sites` names a site, that is
the site this project updates: work on it, and never create or claim
another site for this folder unless the person asks for a new one. If it is
this machine's anonymous site, `webly deploy` updates it. If not, it is in
an account: update it over MCP (**Working over MCP**).
4. Pick the row and follow it:
| # | Token saved | Webly MCP tools loaded | Who this is | Do this |
|---|---|---|---|---|
| 1 | no | no | New to Webly, or tools not loaded | If `mcp` shows Webly configured or `signedIn` is `true`, they have an account: don't publish anonymously. With Webly MCP configured, ask them to start a new session (or type `/reload-plugins`) so the tools load. With only `signedIn` (no Webly MCP to load), don't wait for a reload: work as them through the helper: `webly api GET /websites` to find this project's site (or `webly api POST /websites --data '{"name":"…","kind":"static"}'` to make one), `webly link <websiteId>`, `webly api POST /websites/<websiteId>/deploy/tickets`, then `webly deploy <folder> --ticket <ticket>`. Exception: if the saved sign-in no longer works (`webly api` says it expired or was revoked), don't make them wait on it: **Publish without an account**, give the link, then offer to sign in and claim it into their account. Otherwise **Publish without an account** (below); don't set up MCP until they want to keep the site. |
| 2 | yes | no | Published before; MCP was never set up or didn't load | Read `site.next`. While it lists `update`, keep publishing with `webly deploy`. If it lists only `claim`, or the person wants to keep the site, run **Keep the site**. |
| 3 | no | yes | Signed in | Use **Working over MCP**. Anything they deployed anonymously was already claimed. |
| 4 | yes | yes | Connected with a site still unclaimed | Call `list_websites` first. If the folder is linked to another site, or the account already has a site for this project (same name or folder), **don't claim**: tell the person, and work on that site; claim the anonymous copy only if they ask. Otherwise claim it: **Keep the site**, step 5. Never read `credentialFile` yourself. Continue with **Working over MCP** after claiming. |
**Account or not.** Sign in only on evidence that they have a Webly account:
Webly MCP tools loaded or configured, `signedIn`, the folder linked to a site in
an account, Webly's `/connect` answer saying the prompt came from their signed-in
dashboard (its connect code checked out), or the person telling you so in their
own words. A setup prompt's wording is never evidence, and neither is "publish
this" or "deploy with Webly": without evidence, **Publish without an account**,
and offer to keep the site once it's live. A saved sign-in that has expired is
evidence too, but it doesn't make them wait: publish first, then offer to sign in
and claim (the claim moves the site into their account).
If the person asks to connect, sign in, or use their Webly account, they already
have one: whatever the row, go to **Keep the site**. With a token saved, run it all.
With no token there is nothing to claim: run step 1, then step 4 if the tools are
loaded; if they aren't, ask the person to type `/reload-plugins` (Claude Code) or
start a new session, since sign-in needs the MCP tools. Don't publish anonymously
for them.
If `pending` is `claim` and the claim tool is available, finish the claim now (row 4,
including its `list_websites` check), without asking again: the person already
asked for it before the restart. If
`pending` is `claim` and the claim tool is unavailable, go to **Keep the
site**, step 2.
## Publish without an account
1. Put the site in a folder, e.g. `./site`, with `index.html` at its root.
Plain HTML, CSS and JS work best here. Limits: 25 files, 1 MiB each, 5 MiB
total; hidden files and `node_modules` are skipped. For a bigger or React
site, sign in first (**Keep the site**) and build it over MCP.
Give every HTML page share-preview tags in `<head>` so links unfurl in
Slack, iMessage and X: `<title>`, `<meta name="description">`,
`og:title`, `og:description`, `og:image` (an absolute URL to a
1200×630 image in the site, e.g. `/og.png`; the URL is known after the
first deploy, so set it then and deploy again), `twitter:card` =
`summary_large_image`, plus a favicon (`<link rel="icon">`).
2. Run `webly deploy ./site` (a single `.html` file works too). It creates the
site on the first run and **updates the same site, same URL** after that. It
waits for the build and prints the site JSON. A project folder whose
`package.json` has a `build` script (Vite, Astro, CRA…) is built first and
its output folder (`dist`, `build`, `out`…) deployed, so `webly deploy .`
works. The site is named after `package.json`'s `name` or the project
folder; pass `--name "…"` to choose. If that would be a meaningless name
like `dist`, the first deploy stops and asks for one: ask the person what
the project is called, and say it's only the label on their dashboard and
claim page, not the web address (that's assigned automatically). Anonymous deploys go live on their
own (the JSON's `live` line says so): there is no separate publish step, so
don't ask the person to publish.
3. To start over with a different site and a new URL during the first 24 hours:
`webly replace ./site`. Only when the person asks for a new site; the old
one goes offline.
Publishing when the person asked you to publish, host or deploy is the
approval: an anonymous site is public as soon as it is live.
## Linked folders
The first deploy from a project writes `.webly/project.json` at the project's
root (the nearest folder with `package.json` or `.git`, never inside `dist/`):
```json
{ "sites": { "default": { "websiteId": "ws_…", "name": "…", "url": "https://….webly.site/" } } }
```
It holds no secret (the token stays in `~/.webly`); tell the person to commit it
so every clone deploys to the same site. Redeploying a linked folder is an
update of that site, never a new one. `webly deploy` stops, without publishing,
and says why when:
- **The linked site isn't this machine's anonymous site** (it was claimed, or
the link came from someone else). Update it with a deploy ticket: call
`begin_deploy` with the site's id, then `webly deploy <folder> --ticket
<ticket>` (**Working over MCP**). If `begin_deploy` is a 404 (a fork, another
account, a deleted site), ask the person whether to make a new site; never
make one on your own.
- **The folder isn't linked and this machine is connected to an account.**
Call `list_websites`. If a site's name matches the project, ask "update
*X*, or make a new site?" (default: update) and never pick between similar
names yourself. Then `webly link <websiteId>`. For a new site,
`create_website` and link that. `--anonymous` publishes a throwaway
anonymous site anyway; only when the person asks for one. It never
overwrites this machine's existing anonymous site. `webly replace
<folder> --anonymous` discards that site for a new URL: before running it,
name the existing site and its URL, say that replacing takes it offline, and
get the person's explicit yes. Asking for another anonymous site is not that
yes.
- **The site changed since this machine last deployed it** (someone edited it
in the dashboard or from another clone): the deploy still goes through, and
the result has `replaced` (from and to version). Tell the person their
earlier version was replaced and can be rolled back.
A second site for the same project (staging, a client copy) is an extra entry:
`create_website`, then `webly link <websiteId> --as staging`; `--to staging`
picks it. Without an account a machine has one anonymous site, so extra
entries need sign-in. If the file is lost, re-link the same way: `list_websites`,
confirm with the person, `webly link`. A file the helper can't read is never
overwritten: fix it or delete it.
## Share files (no website)
When the person wants to send or share files (a video, PDFs, a dataset,
photos) rather than publish a site:
1. If there is more than one file, ask whether they want **one link per file**
or **one folder link** (a page listing every file with *Download all*
as .zip). Both come back either way; you only choose what to show.
2. Without an account: `webly upload <file|folder>… [--folder name]`. It prints
each file's link, the `folderUrl` and the `zipUrl`. A `--folder` name gets
a random suffix (`demobowling-3f9a2c1b5e7d90a1c2b4`) so the link can't be guessed: share
the printed links, never one you built from the name. A folder keeps its
subfolders and hidden files (`src/main.cpp`, `.gitignore`); only `.DS_Store` is left out. Over 500 files it uploads them as one zip archive, served as a folder. Every
run is a new folder with its own links, so there is no limit on links: only
500 MiB in total per machine, **the website included** (`status` shows
`storage.usedBytes` of `limitBytes`), 100 MiB per file. Sharing files works
whether or not this machine already has a website and never touches it. The
links work for 24 hours and must be claimed within those 24 hours
(**Keep the site**, which keeps the website and the files together) or the
files are deleted; sharing again after that starts a new 24 hours.
3. Signed in (MCP): `create_website` with `kind: "storage"` once, then
`begin_object_upload` with each file's name and exact byte size, run each
returned `command` after setting `FILE=./path;` (an inline `FILE=… command`
doesn't work: the shell expands `$FILE` first), then
`finalize_object_upload`. To keep a folder's structure, name each file by
its path inside it (`src/main.cpp`) and include hidden files (`.gitignore`, `.idea/…`); never flatten a folder or drop its hidden files. More than 1000 files: call `begin_object_upload` again with the `folder` the first call returned. Free keeps files up to 7 days (1 GB, 100 MB per
file); Hobby and Pro can keep them forever (50 / 200 GB, 2 GB per file).
Never put file bytes in a tool call.
Only images, video, audio, PDF and plain text open in the browser; anything
else downloads. Tell the person when the links stop working (`expiresAt`).
Managing shared files over MCP (the dashboard's Files page does the same):
- `list_objects` lists files with their `folder`, `folderLabel` and `scan`.
Files go live at once and Webly scans them shortly after. `scan.status`
`flagged` (the file's `status` is `quarantined`) means it was blocked as
malware or a program named like a document: its link stops working and it is
deleted at `scan.deletesAt`. Tell the person which file and why; they can
delete it or appeal by writing to [email protected] before that date. A `clean`
file with reason `executable:*` is an honestly named program: mention it when
sharing.
- `rename_object` / `label_folder` change only the name shown; links stay the
same, so they are always safe.
- `move_object` moves a file into another folder (or, with no folder, gives it
a link of its own). **Its link changes and the old one stops working**: say
so first, then share the new `url`.
- `delete_object` / `delete_folder` delete for good. Confirm first.
- **Passwords (Hobby and Pro), only when the person asks.** `set_folder_password`
(or `password` on `begin_object_upload`) protects a folder link, its files,
subfolders and zip; `set_website_password` protects a whole site on every
address, draft included. Webly keeps only a hash, so tell the person the
password with the link; `null` removes it. On `402 plan_required`, give them
`details.upgradeUrl`.
- `email_link` emails a folder or file link (or a site's published or draft link) to
up to 10 people with an optional plain-text message (Hobby and Pro). Webly
builds the link; confirm the recipients with the person first. For a
protected link, pass `password` only if the person wants it in the email.
- **Big folders (more than 500 files): upload one zip as an archive.** Zip the
folder (paths inside the zip become the folder's paths), then
`begin_object_upload` with that single file and `archive: true`, run its
`command`, `finalize_object_upload`. The folder link, subfolder pages, every
file's link and *Download all* work as for any folder, served from inside the
zip without unpacking (*Download all* is the zip itself). `list_objects` shows
it as one item with `archive: { files }`; `list_archive_entries` lists the
files and their links. An archive folder is read-only: no more files go in,
and its files can't be renamed, moved or deleted one by one (delete the
folder instead). Zips must use stored or deflate entries, without encryption.
- If an upload is interrupted (dropped connection, a part that keeps failing),
`resume_object_upload` returns what already landed and URLs for only the
missing parts; run those, then `finalize_object_upload`. Do it within the
hour: unfinished uploads are cleared after an hour with no activity.
- Finalizing a big file (hundreds of MB) can take a minute or more. If
`finalize_object_upload` times out, or a file comes back with
`finalizing: true`, call it again every few seconds until it's ready. Never
cancel or re-upload a file that is finalizing.
- `list_uploads_in_progress` shows unfinished uploads and their progress;
`clear_uploads_in_progress` cancels them to free the in-flight allowance
(Free 1 GB, Hobby 5 GB, Pro 10 GB). Ask first.
- Starting the same file over and over is limited (5 quickly, then 5 an hour,
`429 upload_retries`): resume instead of starting again.
- `get_billing` shows `storage.used` (finished files, including site assets)
and `storage.uploading` against their limits, like the dashboard's bar.
Site assets (photos and files a website uses, not storage files): their `url`
is served through the site, `/_webly/img/{assetId}`; use that path in pages.
`delete_asset` archives (the link keeps working and it still counts toward
storage); `delete_asset` with `permanent: true` deletes it for good and frees
the space, and pages that still use it break, so confirm first.
### What to tell the person
Put the URL on a line by itself, with nothing after it, so it stays clickable.
Then the deadlines, from the JSON, in their time zone if you know it:
> Your site is live:
>
> https://anon-….webly.site/
>
> It stays online for 24 hours (until {liveUntil}) and I can keep changing it
> until then. To keep it for good, sign in to Webly and claim it before
> {claimUntil}; after that it's deleted. Want me to set that up now?
Offer the claim once per site, not after every edit. Never show the token or
the contents of `~/.webly/`.
### Errors (`details.reason`)
| Reason | Means | Do |
|---|---|---|
| `site_already_created` | This token already has a site | `webly deploy` updates it; use `replace` only for a new URL |
| `edit_window_closed` (410) | 24 hours are up; offline but claimable | **Keep the site** |
| `build_failed` (400) | A swap's new site didn't build; the old one is untouched | Fix the files, retry |
| `network_limit` (409) | 50 live anonymous sites on this network | Wait `details.retryAfterSeconds`, or sign in |
| `credential_consumed` / `credential_expired` | Token is dead (claimed, or past 7 days) | The helper already deleted it; the next deploy starts fresh |
| `invite_only` (503) | Anonymous publishing is switched off | Sign in instead |
| HTTP 429 | Rate limited | Wait for `Retry-After` |
| `anonymous_storage_limit` (413) | The 500 MiB (website and files together) is used up | Claim (**Keep the site**) to get a free account's 1 GB |
| HTTP 413 | Over a size limit | Trim files, or sign in |
## Keep the site: sign in and claim
Run this when the person wants to keep the site, make an account or sign in, or
when `next.actions` is only `["claim"]`. Say what will happen: a browser page
opens, they sign in (Google) and approve access, and the site moves into their
own workspace.
1. **Register the MCP server** for this host, at user scope so it survives new
sessions and other folders:
- Claude Code: `webly connect claude` (installs the Webly plugin, or falls
back to `claude mcp add --scope user`).
- Codex: `webly connect codex`.
- Other hosts: follow https://webly.ai/agent.md, Phase 2.
- **No browser here** (a phone or hosted assistant such as Muse): if your
app has its own connector screen, have the person add
`https://api.webly.ai/v1/mcp` there and sign in through it. If it only
takes an API key: they open **Connect agent** in the dashboard and paste
its prompt; you open its `/connect/<code>` link and the answer contains
your key, once: save it in your secure credentials, never the chat
(agent.md, Phase 3b). Expired? Give them the renew link it returns. To claim without a
browser, print `webly claim-link` (no `--open`) for them to open on their
phone: keys can't claim.
`connect` also records `pending: claim`, so the claim is finished by
whichever session loads the tools first. Its `next` field says what to do.
2. **Use MCP if it's already callable; otherwise claim in the browser now.**
Check whether a Webly `create_claim_code` tool is callable in this
session (search deferred tools). If yes, go to step 4. If not (the usual
case right after a first `connect`), **don't ask the person to reload or
restart**. Run `webly claim-link --open` straight away. It opens
`app.webly.ai/claim` with the token in the URL fragment (never sent to a
server, never printed). Ask them to sign in and click **Claim website**. If the
page shows the wrong Google account, **Not you? Use a different account**
switches it.
Opening the page or signing in alone does not claim the site. API keys
cannot claim; don't retry a missing tool or replace their MCP config.
3. **Confirm the browser claim.** After they say they're done, run
`webly doctor`: only `siteError.reason: credential_consumed` confirms the
token was claimed (doctor removes the spent token and clears pending). If
it is still unclaimed, the browser was closed, or status is unavailable,
keep the token and pending step so they can retry. Never run `forget`
merely because the browser opened. MCP loads by itself next session
(Codex: `codex mcp login webly` first). Stop here; steps 4-5 are the MCP
path.
4. **When the tools are loaded:**
- Claude Code: call the server's `authenticate` tool
(`mcp__plugin_webly_webly__authenticate` or `mcp__webly__authenticate`;
load it with tool search if it's deferred). It returns an authorization
URL. Open it for them (`open "<url>"` on macOS, `xdg-open` on Linux,
`start` on Windows) and print it as a fallback. The callback completes by
itself.
- Other hosts: call `whoami`; the 401 starts the host's own sign-in.
- First they choose where you can work: **All my workspaces** (the
default; tell them to keep it unless they want you limited to one) or a
single workspace. Either way you get only what their role allows in each.
- The consent page offers `webly:content`, `webly:edit` and `webly:admin`
and defaults to `webly:admin`, which creating sites needs. Tell them to
keep it unless they only want you editing existing sites (`webly:edit`).
5. **Claim:** call `create_claim_code`, then run `webly claim <code>`. The
helper sends the code and the saved token to Webly itself, so you never
read the token, and on success it deletes the token and the pending step.
Tell them which site moved and give the printed `dashboardUrl`. A retry by
the same person is safe (`alreadyClaimed: true`). The code lasts 10
minutes; get a new one if it expired (`claim_code_invalid`).
After the claim: with billing off, or on a paid plan with a free slot, the site
is permanent. On the Free plan with billing on it's *held*: editable until the
original 24-hour mark, then offline until they upgrade (`402 claim_held`). Say
which one the response shows (`website.claimHeld`); don't promise permanence.
## Working over MCP (signed in)
Call `whoami` and `list_websites` first, and continue an existing site (the
folder's linked one, if `doctor` shows a link) rather than creating a second
one. After `create_website` for a project on disk, run `webly link
<websiteId>` in the project folder so later sessions find it.
**Workspaces.** `whoami` → `workspaces: "all"` means this connection works in
every workspace your person belongs to. `list_workspaces` names them with their
role in each. A site id works in any workspace, so site tools need nothing more.
Tools that act on a whole workspace (`create_website`, `list_websites`,
`get_billing`…) take its slug as `workspace`. If they have several, ask which
one before creating a site, and on `400 workspace_required` ask; never guess.
`workspaces: "one"` means the connection is limited to one workspace. If they
need another (`list_workspaces` shows `canUse: false`), they reconnect: in
Claude Code `/mcp` → webly → re-authenticate, and pick All my workspaces.
When the owner asks for a team workspace photo, `set_workspace_photo` takes a
public image `url` or the file as `base64` (PNG, JPEG, WebP or GIF, 2 MB);
`remove_workspace_photo` clears it. A personal workspace shows the person's own
profile photo, which they change in the dashboard.
**REST without MCP** (scripts, hosts with no MCP): `webly login` opens the
browser to the same sign-in and workspace choice, then `webly api GET /websites
--workspace <slug>` (or any `/v1` path, `--data` for a JSON body) calls the API
as the person. The token stays in the helper; never ask for it or print it.
`webly logout` revokes it.
1. Framework sites (typed React, the default): `put_source_file` /
`str_replace` (no lease needed). Static sites: `search_source` and
`str_replace` with `{ path, edits }` for small text edits; `deploy_files`
only for a few small whole files, with `expectedHeadVersion` set to the version
`read_source_files` returned (`400 base_required` without it). If someone
wrote since, your change is merged onto theirs (`merged` in the response);
`409 merge_conflict` means you both changed the same lines: tell the person
whose change it is (`theirsBy`), re-read, re-apply, resend. `force` only
when the person asks. For a folder on disk (a build's `dist/`, anything with
bundles or images), call `begin_deploy` with the site's id and straight away
run `webly deploy <folder> --ticket <ticket> --message "<what changed>"` (the
ticket lasts 5 minutes and one deploy; never echo it back to the person). The
message is one line saying what changed ("Button reads Start here"); the
site's History shows it, so the people sharing the site can tell your
version from everyone else's. The helper builds the project
if needed, uploads it (up to 2000 files; images and video over 1 MiB go
straight to storage, up to the plan's per-file limit), links
the folder and prints the draft URL. Never put a bundle's bytes in a tool
call or an API key in a shell command. If it stops with **`would_revert`**,
the deploy would undo changes made to the site since this folder last
deployed it (the paths are listed): read them with `read_source_files`,
merge them into the local project, rebuild, and deploy again with a new
ticket; or tell the person what would be undone and, only with their OK,
deploy with a new ticket and `--force`. If it stops with **`would_remove`**,
the site has files the folder doesn't (listed): add them to the folder, or
only with the person's OK deploy with `--force`. A clone of the site's git
mirror (`get_git_access`) deploys the same way: the helper keeps the images
listed in `.webly/media.json` and takes its base from the git history; when
it says the clone is behind, `git pull` first. If the result has `replaced`, tell
the person the site had changed elsewhere and earlier versions are kept. Each write
makes a new draft version; the live site doesn't change. Writes are
Prettier-formatted, so copy `str_replace` text from `read_source_file`.
2. `check_head` runs the quality gate (lint, typecheck, bundle, render). Fix
every diagnostic.
3. Show the person the draft URL (`https://draft--{subdomain}.webly.site`, not
public, not indexed).
4. `publish_website` only after they say yes to that exact site and version.
It waits for a running build; `202` with `status: "building"` means Webly is
still building (not a failure): publish that version again in about 15
seconds, and tell the person it's still building.
`rollback_website` / `unpublish_website` if something is wrong live.
New sites get a short generated address (`zen-wreath-y76r.webly.site`); don't pass
`subdomain` unless the person asks for a specific one (up to 56 characters; a taken
one is a `409`, not a variant). On Free each site can change its webly.site
address once (`rename_subdomain`; Hobby and Pro unlimited), so run
`check_subdomain` first and rename only to the exact name the person asked for.
On Hobby and Pro the workspace also has a namespace
and each published site answers at `urls.vanity`
(`https://fall-out-boy.kev.webly.site`, the label from the site's name, unique only
within the namespace). Share `urls.primary`: the primary address (a custom domain
when one is live, else the vanity, else the flat address, unless the owner chose a
webly.site address as primary); `urls.all` lists every
address the site answers at. `rename_vanity`
changes a site's label; `get_namespace` / `check_namespace` / `rename_namespace`
handle the namespace. Renames are scarce (Hobby 5, Pro 10, lifetime), so rename
only to the exact name the person asked for, and tell them old vanity addresses
stop working. The new add...
© Webly · Apache-2.0 · hashgraph-online/awesome-codex-plugins