create-curb.js 1.16.0-preview.1 → 1.16.0-preview.11

package/README.md

@@ -1,14 +1,57 @@
-# create-curb.js
+<div align="center">
+  <a href="https://curbjs.com"><img alt="Curb" src="https://curbjs.com/og.png?v=2" width="360"></a>
+  <h1>create-curb.js</h1>
+  <p><strong>The framework for production voice agents.</strong></p>
+  <p>Dynamic prompting with full control, in plain code you own. Your coding agent can build, test, and simulate conversations before it ships.</p>
+  <a href="https://www.npmjs.com/package/create-curb.js"><img alt="npm version" src="https://img.shields.io/npm/v/create-curb.js.svg?style=for-the-badge&labelColor=000000"></a>
+  <a href="https://curbjs.com/docs/installation/"><img alt="Node.js version" src="https://img.shields.io/node/v/create-curb.js.svg?style=for-the-badge&labelColor=000000"></a>
+  <a href="https://www.apache.org/licenses/LICENSE-2.0"><img alt="License" src="https://img.shields.io/npm/l/create-curb.js.svg?style=for-the-badge&labelColor=000000"></a>
+</div>
 
-Create a voice agent on KugelAudio:
+`create-curb.js` creates a Curb voice agent in one command. You need Node.js 22.12 or newer.
 
-```sh
+```bash
 npm create curb.js my-agent
 cd my-agent
-npx curb login
-npm run dev
+npx curb login         # sign in once in the browser
+npm run dev            # open http://127.0.0.1:4000
 ```
 
-`npx curb login` signs you in once in your browser. `npm run dev` gives you a link to talk to your agent.
+`npx curb login` signs you in once in your browser. `npm run dev` gives you a link to talk to your
+agent. Run `npm create curb.js .` to create the project in an empty folder instead. `npm create
+curb` is the same installer under a shorter name.
+
+The framework's package is `curb.js`, not `curb`, which is unrelated. The project depends on it
+through an alias, `"curb": "npm:curb.js@<version>"`, so your code imports from `"curb"`.
+
+## Providers
+
+The defaults run speech-to-text, the LLM and text-to-speech on KugelAudio with one sign-in. Nothing
+is asked: flags after `--` pick other providers, and their API keys are read from the environment.
+
+```bash
+npm create curb.js my-agent -- --llm anthropic --tts elevenlabs
+```
+
+| Flag | Providers |
+| --- | --- |
+| `--stt` | `kugelaudio`, `azure`, `deepgram`, `elevenlabs`, `openai` |
+| `--llm` | `kugelaudio`, `anthropic`, `openai`, `openai-compatible` |
+| `--tts` | `kugelaudio`, `cartesia`, `elevenlabs`, `openai` |
+
+`npm create curb.js -- --help` lists every option.
+
+## Previews
 
 To try the newest preview, create the project with `npm create curb.js@preview my-agent`.
+
+## Documentation
+
+- [Installation](https://curbjs.com/docs/installation/): create a project and talk to it.
+- [Project structure](https://curbjs.com/docs/project-structure/): what the installer created.
+- [All docs](https://curbjs.com/docs/) on curbjs.com.
+
+## License
+
+Licensed under the [Apache License 2.0](https://www.apache.org/licenses/LICENSE-2.0).
+Copyright 2026 KugelAudio GmbH. Built by [KugelAudio](https://kugelaudio.com).

package/RELEASE_NOTES.md

@@ -1,5 +1,15 @@
 
 ### Added
 
+- Make npm create curb work as the short form (#49)
 - Make npm the only way to get and update Curb (#36)
 
+
+### Fixed
+
+- Refuse unknown arguments to curb upgrade (#39)
+- Explain a target folder that cannot be created (#40)
+- Say what hosted deploys accept today (#46)
+- Correct what the help and guide say about conversations and build (#41)
+- Start curb outside a project from the npm package (#38)
+

package/bin/create-curb.js

@@ -12,6 +12,7 @@ import {
   readFileSync,
   readdirSync,
   renameSync,
+  statSync,
   writeFileSync,
 } from "node:fs";
 import { basename, dirname, join, resolve } from "node:path";
@@ -118,6 +119,9 @@ if (!/^[a-z0-9][a-z0-9._-]*$/.test(appName)) {
     `"${appName}" is not a valid project name — use lowercase letters, digits, ".", "_", "-" (npm package name rules).`,
   );
 }
+if (existsSync(targetDir) && !statSync(targetDir).isDirectory()) {
+  fail(`${targetDir} is a file. Choose a new directory.`);
+}
 if (existsSync(targetDir) && readdirSync(targetDir).length > 0) {
   fail(`${targetDir} already exists and is not empty. Choose a new directory.`);
 }
@@ -301,8 +305,12 @@ const envValues = finalizeEnv();
 
 // -------------------------------------------------------------- copy template
 const templateDir = resolve(dirname(fileURLToPath(import.meta.url)), "../templates/default");
-mkdirSync(targetDir, { recursive: true });
-cpSync(templateDir, targetDir, { recursive: true });
+try {
+  mkdirSync(targetDir, { recursive: true });
+  cpSync(templateDir, targetDir, { recursive: true });
+} catch (error) {
+  fail(`Cannot create the project in ${targetDir} (${error.code ?? error.message}). Choose a directory you can write to.`);
+}
 
 // npm strips .gitignore from published tarballs, so the template ships it without the
 // leading dot. The project carries no .npmrc: hosted builds refuse one.

package/package.json

@@ -1,12 +1,34 @@
 {
   "name": "create-curb.js",
-  "version": "1.16.0-preview.1",
-  "description": "Create a Curb voice agent: npm create curb.js my-agent",
+  "version": "1.16.0-preview.11",
+  "description": "Create a Curb voice agent with one command. The framework for production voice agents.",
+  "keywords": [
+    "curb",
+    "curb.js",
+    "scaffold",
+    "starter",
+    "template",
+    "cli",
+    "voice-agent",
+    "voice-ai",
+    "agent-framework",
+    "conversational-ai",
+    "react",
+    "typescript"
+  ],
   "license": "Apache-2.0",
+  "author": {
+    "name": "KugelAudio GmbH",
+    "url": "https://kugelaudio.com"
+  },
   "homepage": "https://curbjs.com",
   "repository": {
     "type": "git",
-    "url": "git+https://github.com/Kugelaudio/curb.js.git"
+    "url": "git+https://github.com/Kugelaudio/curb.js.git",
+    "directory": "packages/create-curb.js"
+  },
+  "bugs": {
+    "url": "https://curbjs.com/docs/reporting-a-bug/"
   },
   "type": "module",
   "engines": {

package/templates/default/AGENTS.md

@@ -68,8 +68,8 @@ A fault in Curb itself is reported as `reporting-a-bug.md` says, with the person
 4. `npm run dev`: the browser preview with chat and voice on http://127.0.0.1:4000. Open it in the person's own browser yourself (`open http://127.0.0.1:4000` on macOS, `xdg-open` on Linux, `start` on Windows): a browser built into an app, like the Claude app's Browser pane, blocks the microphone, so a call there fails.
 
 **To improve the agent, read its recent conversations first.** `npx curb conversations`
-prints the last ones the preview and `curb say` ran, with every tool call, its
-input and its result (`--last <n>`, `--id <id>`, `--json`). Calls to the hosted agent are
+prints the last ones the preview ran, with every tool call, its input and its
+result (`--last <n>`, `--id <id>`, `--json`). Calls to the hosted agent are
 read with the `kugelaudio` MCP server's `agent_list_calls`, `agent_get_call` and
 `agent_get_call_timeline` where it is signed in. Find where the agent went wrong, fix that,
 and turn the conversation into a scenario so it stays fixed.
@@ -90,7 +90,7 @@ src/agent/agent.tsx         the agent, where every conversation starts: <Agent l
 src/agent/<name>/skill.tsx  one folder per skill, opened when needed; nested folders are sub-skills
 context/*.md                knowledge a skill reads with <Text src="file.md" />
 .env                        other providers' keys, an optional KUGELAUDIO_API_KEY; never commit it
-.data/conversations.jsonl   every local conversation: tool calls, inputs, results
+.data/conversations.jsonl   every preview conversation: tool calls, inputs, results
 ```
 
 ## Rules
@@ -128,12 +128,18 @@ context/*.md                knowledge a skill reads with <Text src="file.md" />
 - **What counts as a successful call is code too**: `defineAgent({ goals, experiments })`
   names each goal (a tool that must succeed, or a sentence a judge checks) and each A/B test
   with its weighted variants, and `useVariant("<test>")` reads the call's variant. The
-  dashboard only shows the results; `authoring.md` ("Goals and A/B experiments") has the rules.
+  dashboard only shows the results; `goals-and-experiments.md` has the rules.
 - Tool names are unique snake_case verbs. Zod `.describe()` texts and error messages are
   read by the model, so phrase them as the next step.
 
 ## Deploy to KugelAudio
 
+**Hosted deploys do not accept a project that installs Curb from npm yet, and this project
+does.** Tell the user so before you try to deploy, and do not change the `curb` dependency to
+get around it. Today the project runs locally with `npm run dev` and goes live on the user's own
+servers (`node_modules/curb/docs/agent-guide/hosting.md`). The rest of this section applies once
+hosted deploys accept the project.
+
 `.mcp.json` registers two MCP servers for Claude Code started in this folder (it asks once
 to use them): `kugelaudio`, your KugelAudio account (sign in from `/mcp`), and
 `kugelaudio-docs`, the docs (no sign-in). Another client adds

package/templates/default/README.md

@@ -41,6 +41,10 @@ Type or talk in the preview:
 
 ## Deploy to KugelAudio
 
+Hosted deploys do not accept a project that installs Curb from npm yet, and this project does.
+Today it runs locally with `npm run dev` and goes live on your own servers: see
+`node_modules/curb/docs/agent-guide/hosting.md`. Once hosted deploys accept it:
+
 ```bash
 npx curb login
 npx curb secrets push
@@ -49,8 +53,7 @@ npx curb deploy
 
 `deploy` prints the playground URL where you can call the live agent. Your `.env` is
 never uploaded; `secrets push` sends its values and prints only their names. Details:
-`node_modules/curb/docs/agent-guide/deploying.md`. To run the voice worker on your own LiveKit instead,
-see `node_modules/curb/docs/agent-guide/hosting.md`.
+`node_modules/curb/docs/agent-guide/deploying.md`.
 
 ## Learn more