# Picflow MCP server through Cotlet.dev

Connect Picflow to ChatGPT, Claude, Codex and Claude Code through Cotlet.dev. Create client galleries, summarize feedback and prepare replies in your own signed-in Picflow workspace.

This is an unofficial integration, not affiliated with Picflow. Cotlet.dev hosts the connection through your Cloud Browser. This repository provides an optional Python stdio MCP bridge.

## Connect your account

1. [Request access to Picflow](https://cotlet.dev/signup?app=picflow).
2. Open https://app.picflow.com/ in your Cotlet.dev Cloud Browser, sign in yourself and enable Picflow for agents.
3. Connect your agent. Refresh the connector if you enabled Picflow after connecting.

### ChatGPT

Open Settings > Apps & Connectors > Create. Enter `https://cotlet.dev/mcp/v1/connect` as the MCP server URL and choose OAuth.

### Claude

Open Settings > Connectors > Add custom connector and enter `https://cotlet.dev/mcp/v1/connect`.

### Codex

Create a Cotlet.dev API key in your console and set `COTLET_API_KEY` in your environment. Add:

```toml
[mcp_servers.cotlet]
url = "https://cotlet.dev/mcp/v1"
bearer_token_env_var = "COTLET_API_KEY"
```

### Claude Code

```bash
claude mcp add --transport http cotlet https://cotlet.dev/mcp/v1 --header "Authorization: Bearer $COTLET_API_KEY"
```

## Gallery and proofing workflows

Examples based on the [live Picflow integration reference](https://cotlet.dev/integrations/picflow-api-mcp-server):

- Create a gallery from a Dropbox, Google Drive or Box folder.
- Summarize favorites, comments, annotations and approval flags.
- Draft replies to reviewers for approval before posting.
- Prepare retouch annotations and set asset statuses.
- Configure Approve-Only, an approval limit and download settings.
- Protect gallery access and organize delivered galleries into folders.

The live reference lists 13 capabilities, including 2 read and 11 write capabilities. The hosted groups are `picflow_read` and `picflow_write`. Operations:

- `upload_assets`
- `import_from_cloud_storage`
- `manage_folders`
- `move_copy_gallery`
- `manage_gallery_sections`
- `manage_gallery_access`
- `get_share_links`
- `read_feedback`
- `reply_to_comment`
- `add_annotation`
- `set_asset_status`
- `configure_review_settings`
- `manage_team_members`

`get_share_links` and `read_feedback` only read. Writes ask for confirmation by default; confirmation can be changed per tool. Review replies, access changes, invitations and deletions before execution. Deleting a folder deletes its galleries. Team writes require an owner or admin login. Features depend on your Picflow plan and permissions; first-time cloud storage authorization is completed by you in the Cloud Browser.

## Optional local stdio bridge

Python 3.10 or newer is required, with no runtime dependencies. Clone this repository and configure a client that supports stdio:

```json
{"mcpServers":{"picflow":{"command":"python3","args":["/absolute/path/to/picflow-mcp/picflow_agent.py"],"env":{"COTLET_API_KEY":"your-cotlet-key"}}}}
```

The local bridge lists seven browser helpers: `picflow_guide`, `picflow_status`, `picflow_open`, `picflow_read`, `picflow_click`, `picflow_type` and `picflow_close`. They forward to Cotlet.dev's shared browser tools and are distinct from the 13 hosted workflow capabilities. Call the guide first and close the browser when finished. Discovery works without a key; executing a tool requires your Cotlet.dev key and connected account.

```bash
python3 picflow_agent.py connect
python3 picflow_agent.py guide
python3 picflow_agent.py status
python3 picflow_agent.py serve
```

You can also build the included Dockerfile and pass `COTLET_API_KEY` as an environment variable.

## Reference and access

- [Picflow connection guide and operation reference](https://cotlet.dev/integrations/picflow-api-mcp-server?utm_source=github&utm_medium=readme&utm_campaign=picflow)
- [Request access](https://cotlet.dev/signup?app=picflow)
- [Cotlet.dev console](https://cotlet.dev/app)

Licensed under Elastic License 2.0. See LICENSE.