# Schedule a Week of Social Posts in Buffer Yes. Given an article, an agent writes five posts per connected Buffer channel and saves them as drafts over the next 7 days for you to review. [Run this on Rush](https://cloud.getrush.ai/run/schedule-a-week-of-posts/buffer/). ## What the agent does The run uses two inputs: the article to post about, and your Buffer account. The request below names the article by URL and the agent opens it in the sandbox's browser; an article that is not public goes in as a `document` block in `input` instead. Buffer is reached through Buffer's remote MCP server, `https://mcp.buffer.com/mcp`, the Buffer entry in the Rush connector catalog, connected once with `POST /me/connections {"type": "mcp", "name": "buffer", "url": "https://mcp.buffer.com/mcp", "auth": {"mode": "oauth"}}`. The agent reads the article and picks out its points: the claim in the headline, each section's main idea, a number or a quote worth repeating. It needs five distinct points, so a short changelog may give fewer, in which case it says so instead of padding. Then it lists your Buffer channels and reads each one's recent posts, to see how long they run, whether they use hashtags and which times of day you post. For each channel it writes five posts, one point each, sized for the network: short for X and Threads, longer with a line break or two for LinkedIn, the link placed where each network shows it. It matches the voice of the channel's past posts rather than one house style. Every post links back to the article. Last, it adds each post to Buffer as a draft, with a time on one of the next 7 days at a slot the channel usually posts, no two posts on one channel within a few hours. It never sends a post to the queue for publishing and never publishes now; the prompt says drafts only, and the drafts wait in Buffer for a person to approve. ## Start it with one API call `POST https://api.prix.dev/o/{org}/sessions` ```json { "agent": "claude", "input": "Read the article at https://example.com/blog/launch-post. Use the Buffer connection to list the connected channels and read each one's recent posts for length, tone, hashtags and usual posting times. Write five posts for each channel, each making one different point from the article and linking to it, sized and written for that network. Add them to Buffer as drafts spread across the next 7 days at the channel's usual posting times. Do not publish or queue anything for publishing; drafts only. In your answer, list every draft with its channel and time.", "connections": [ "buffer" ] } ``` 1. **Get an API key.** In the console, open Settings, then API keys, and create a key for your org. Send it in the Authorization header as a Bearer token. 2. **Send the request.** POST the body in request.json to https://api.prix.dev/o/{org}/sessions, with your repository as the resources URL. The run starts in its own cloud sandbox. 3. **Read what came back.** Follow the run on its live stream or wait for the run.completed webhook, then read the result and its trajectory. ## What comes back What comes back is in two places. In Buffer, each channel has five new drafts with times across the coming week. In the session, the agent's `result` lists every draft with its channel, its planned time and its text, and says which point from the article each one makes, so a reviewer can read the whole week in one place before opening Buffer. The review is the point of drafts. Check that every claim in a post is in the article (the agent writes from the article, but a short post can overstate it), that links resolve, and that times do not collide with something already scheduled. Approve, edit or delete each draft in Buffer. The transcript has every Buffer call the agent made, which shows that nothing was published. ## Run it on every new blog post Tie the run to publishing. Your blog's RSS feed or CMS can call a small endpoint when a post goes live, and that endpoint sends this request with the new post's URL in the prompt. A changelog page works the same way on each release. The `run.completed` webhook tells you the drafts are in Buffer. Post the session's `result` to the channel where your team approves social posts, and the week's drafts are reviewed the same day the article ships. ## Why run it on Rush? Turning one article into a week of posts for four networks is a couple of hours of rewriting, and it is the step small teams skip. The agent reads the article and your past posts in its own sandbox, writes for each network, and puts the drafts where you already schedule, so the work left is approval. Nothing goes out on its own: the prompt says drafts only, and the transcript records every call to Buffer. A team with several brands, each with its own Buffer connection, can run one session per brand at the same time. ## Frequently asked questions ### Will it publish anything? No. It saves drafts with a planned time and the prompt rules out publishing. A person approves each draft in Buffer before it goes out. ### Can I give it an article that is not public yet? Yes. Send `input` as a list of blocks with the prompt as text and the article as a `document` block, or as a file uploaded through `POST /o/{org}/files`. The agent reads it from `/workspace/.rush/attachments/`. ### How does it pick posting times? From when each channel's recent posts went out. Name times in the prompt if you want specific slots instead. ### Can it write fewer posts for some channels? Yes. Say how many per channel in the prompt, for example three for LinkedIn and five for X. A short article may also give fewer distinct points than five, and the agent says so.