---
title: "wendy chat"
description: "Build apps and work with your devices through a conversation"
---

## Syntax

```sh
wendy chat [prompt...] [flags]
```

## Examples

```sh
  wendy chat
  wendy chat "Help me get this project running"
  wendy chat --setup
  wendy chat -C ./my-app
  wendy chat --prompt "Which of my devices are online?"
  wendy chat --prompt "Why does app A on device X fail?" --json --yes
```

## Arguments

Argument notation from the CLI: `[prompt...]`. Angle brackets mark required values; square brackets mark optional values. See Details for command-specific validation.

## Flags

Flags below include inherited options. Types and defaults come from the CLI source. An empty default means no value is configured.

| Flag | Type | Default | Description |
| --- | --- | --- | --- |
| `--device` / `-d` | string | (empty) | Preferred Wendy device (you can also choose while chatting) |
| `--directory` / `-C` | string | `.` | Project directory for local file and command tools |
| `--help-all` | bool | `false` | Show advanced connection options |
| `--json` | bool | `false` | Output in JSON format |
| `--list-profiles` | bool | `false` | List available agent profiles and exit |
| `--no-memory` | bool | `false` | Disable recall and learning of local chat memories |
| `--profile` | string | `general` | Agent specialization; see --list-profiles |
| `--prompt` / `-p` | string | (empty) | Run one turn without the chat screen and exit; combine with --json and --yes |
| `--setup` | bool | `false` | Choose or change your AI and model |
| `--voice` | bool | `false` | Listen and speak with GPT Live (or use /voice in chat) |

## Output

Output depends on the operation and selected target. Use the command help for additional guidance..

## JSON output

The global `--json` option requests machine-readable output. The CLI also selects JSON mode when no interactive terminal is attached. Support and payloads are command-specific; this flag does not guarantee a single JSON document for commands that stream logs or invoke build tools. Consult Details before parsing command output.

## Exit codes

The CLI entry point returns `0` on success and for its user-cancelled or default-cleared results. Other returned errors exit with `1` and print a diagnostic to stderr. Some operations may terminate through subprocess or signal handling; do not infer application health from the CLI exit code alone.

## Common errors

Check argument values and flag combinations against the synopsis above. For discovery, build, port, and readiness failures, see [deployment troubleshooting](../../../errors/). Command-specific constraints are in Details.

## See also

- [wendy](../../../reference/cli/)
- [CLI conventions and configuration](../../../reference/)

## Details

Chat with Wendy to build apps, deploy them, and work with your devices.

Just run wendy chat. I'll help you choose a local AI or a cloud service,
connect it, and pick a model. Your choice is remembered for next time.

Use --setup whenever you want to change your AI or model.

Scripts and other programs can run one turn without the chat screen with
--prompt. Add --json for one event per line; add --yes to allow tools that
would otherwise ask a person.

Use the command help for additional guidance.