> ## Documentation index > Fetch the complete documentation index at: https://runstanza.com/docs/llms.txt > Ask it with GET https://runstanza.com/docs/ask?q=. Short keywords rank best. > Use these to discover all available pages before exploring further. # Connect your agent > Connect ChatGPT, Claude, or another MCP client to your Context, drafts, and articles. Stanza connects to the agent you already write with over MCP. Once it's connected, your agent can read your brand guide and references, add references, create drafts, edit articles, and comment on brand pages and articles. In ChatGPT and Claude, you can also open the Stanza editor right in the conversation. You need a Stanza account with access to a project. If you don't have one, [create an account](https://app.runstanza.com/signup) and follow the [quickstart](/quickstart). The quickest way to connect is from Stanza: choose **Connect your agent** in the Editor, pick your agent, and follow the steps. Stanza notices when the connection is made and gives you a first prompt to try. ## Connection details | Field | Value | | --- | --- | | Name | `Stanza` | | MCP server URL | `https://app.runstanza.com/api/mcp` | | Authentication | OAuth | During setup, you sign in on Stanza's own page and choose **Allow**. You don't need an API key or client secret. Each person connects with their own Stanza account. ## Choose your agent Add Stanza as a custom MCP plugin. Add Stanza as a custom connector. Connect any client that supports remote MCP servers with OAuth. ## What you're allowing Stanza's sign-in page lists the permissions your agent asks for. You allow them together. | Permission | What it allows | | --- | --- | | `read` | See the brands you belong to and read each brand's Context. | | `write` | Save drafts, add references, and leave comments in those brands. | | `edit` | Edit existing articles. | | `offline_access` | Stay connected without asking you to sign in again. | The connection reaches every project your account belongs to, including ones you join later. Stanza checks your access on every request, so removing someone from a project stops their agent from reaching it. ## What your agent can do | Tool | What it does | | --- | --- | | `list_organizations` | List your organizations and the projects you can open in each. | | `get_context` | Show a project's Context: its brand pages, references, and articles. | | `search_context` | Find relevant brand pages, references, and articles. | | `read_context_file` | Read a brand page, reference, or article. | | `get_voice` | Get the project's voice guidance. | | `list_skills` | List the project's writing skills from the **Skills** tab. | | `read_skill` | Read a skill's instructions so the agent can follow them. | | `add_reference` | Add a web page or pasted text to the project's references. | | `comment` | Comment on a passage of a brand page or article. | | `create_draft` | Create a new draft and return a link to it in Stanza. | | `read_article` | Read an article's current text and open comments. | | `list_comments` | Read an article's open comments. | | `edit_article` | Make targeted edits to an existing article. | | `list_stanza_projects_and_articles` | List projects and a project's articles. | | `open_stanza_editor` | Open the Stanza editor in ChatGPT or Claude. | A few things to know: - **Agents don't edit brand pages.** They leave a comment on the passage they'd change. You make the edit. - **You can tell what your agent did.** Its comments and live edits show your name with a small ChatGPT or Claude mark. If you dictate a comment and ask your agent to leave it for you, it shows under your name alone. - **Adding a reference only adds it.** Agents can't change or remove references, and adding one doesn't change your brand pages. - **Drafts and articles are live.** Edits your agent makes show up for everyone with the draft open in Stanza. - **Agents can read your team's comments.** Ask, for example, "Address the comments on my launch draft." ## Try it Check the connection: > Use Stanza to list the projects I can access. Then replace "Acme" with your project's name: > Use the Acme project in Stanza. Read our voice page and relevant references, then draft > a short introduction for new customers. Show which files support the factual claims. When you're happy with it: > Save this draft to the Acme project in Stanza with the title "An introduction to Acme." Open the link your agent returns to keep working on it in the [Editor](/editor). When you find something worth keeping: > Add https://techcrunch.com/2026/10/20/acme-raises to Acme's references in Stanza with > the note "Our funding coverage." Then comment on anything it changes on our > positioning page. Ask your agent to **use Stanza** by name. Being connected doesn't guarantee it will check your Context for every request. ## Troubleshooting | Problem | What to check | | --- | --- | | No projects appear | Make sure you connected with the same Stanza account you use in the app, and that you're a member of the project. | | The agent asks which project to use | Name the project, or ask it to list your projects first. | | Reading works but saving or commenting is refused | Reconnect and allow **Save drafts, add references, and leave comments in those brands**. Also check whether your workspace in ChatGPT or Claude restricts Stanza's tools. | | Existing articles can't be edited | Reconnect and allow **Edit existing articles**. | | Stanza says project tools are paused | An organization owner can see why in Stanza's **Organization settings**. | | Stanza says you've made too many requests | Wait a few minutes and try again. | | You're asked to sign in again often | Reconnect once and allow **Stay connected without asking you to sign in again**. | To stop using Stanza in a chat, turn it off in your agent's tools menu. To remove the connection, use your agent's connector or plugin settings. Signing out of Stanza doesn't end a connection you've already allowed.