Write SEO blog post drafts with GitHub Copilot
Writing blog post drafts with GitHub Copilot means asking Agent mode in VS Code to have StoreSEO write a post into a draft article in your Shopify blog, then tracking the job from the same chat panel.
Last updated: October 2026
Set up GitHub Copilot for StoreSEO
Skip this if GitHub Copilot is already connected. The full guide, with the alternatives, is on the GitHub Copilot page.
- 1
Add StoreSEO to .vscode/mcp.json
Create or open .vscode/mcp.json in your workspace and paste this. VS Code uses a top-level servers key, not mcpServers, and a remote server needs type http.
JSON{ "servers": { "storeseo": { "type": "http", "url": "https://mcp.storeseo.com/mcp" } } } - 2
Start the server and sign in
Run MCP: List Servers from the Command Palette, pick storeseo and start it. If it asks you to sign in, pick your Shopify store in the browser and approve access. Then open Copilot Chat in Agent mode, check the tools icon lists StoreSEO and ask this.
PromptUse StoreSEO to show me my store's SEO score breakdown.
How does this work in GitHub Copilot?
You open Copilot Chat in VS Code, switch to Agent mode and ask for the post: a 1,300 word article on picking a yoga mat for the Guides blog. Agent mode calls storeseo_start and then storeseo_list_blogs, and VS Code lists the tools it used under the reply, so you can expand each call. The server comes from .vscode/mcp.json or your user configuration, and you can see it in MCP: List Servers.
VS Code puts a confirmation in front of MCP tools, and StoreSEO adds its own instruction: the agent states the topic, keyword, length and credit cost, and waits for your yes. For storeseo_generate_blog_post you therefore get a message from Copilot and then the editor's confirmation card, where you can look at the arguments before you continue. The writing runs in the background for a few minutes. Copilot polls storeseo_get_blog_post_job and reports steps completed out of total, and the editor stays usable while it does.
When the job completes, ask Copilot to read the draft with storeseo_get_resource_seo and, if you like, write the findings into a Markdown file in the workspace so they sit next to your code. The post itself is a draft article in Shopify. You edit it and publish it from Shopify admin. On Copilot Business and Enterprise an organization admin must have enabled MCP servers in Copilot policy before any of this works.
-
You: Write a 1,300 word draft on picking a yoga mat for the Guides blog.
- storeseo_start read-only
- storeseo_list_blogs read-only
-
GitHub Copilot: Guides is available. This post will spend AI credit that grows with its length. Shall I start it?
-
You: Yes, start it.
- storeseo_generate_blog_post needs approval
- storeseo_get_blog_post_job read-only
-
GitHub Copilot: The job is running, 3 of 5 steps done. I will poll it and tell you when the draft is ready in Shopify admin.
What to ask GitHub Copilot
Use the StoreSEO tools to list my Shopify blogs and their ids.
Write a 1,300 word draft on picking a yoga mat for the Guides blog. Tell me the AI credit cost and wait for my confirmation before you start.
Check the blog post job you started and, when it completes, save the draft's SEO issues to blog-notes.md in this workspace.
- List my blogs and write a 1,200 word draft on how to care for linen shirts in the News blog. Tell me the cost first.
- Write a draft post targeting best gifts for coffee lovers, with a featured image.
- How far along is the blog post you started?
- Check the SEO score of the draft you just wrote.
- Draft three post ideas for my store and write the strongest one.
What StoreSEO does: blog post drafts
StoreSEO writes a full blog post with AI and saves it as a draft article in the Shopify blog you choose, optionally with a featured image. Nothing is published: you review and publish from Shopify admin.
Requires: Spends AI Content Optimizer credit; the cost grows with word count and a featured image. Blogs must be synced in the StoreSEO app.
- 1
Pick the blog
The agent lists your synced blogs, the sections posts go into, with their ids. An empty list means the blogs need syncing in the StoreSEO app first.
- 2
Agree the brief and cost
Topic, keyword, length and whether to add a featured image. The cost grows with length and the image, so the agent tells you and waits for your yes.
- 3
Generate in the background
Writing takes a few minutes. The agent polls the job and reports progress, and does not present the post as finished while it runs.
- 4
Check the draft
When the job completes, the agent can read the draft article's SEO record. You review and publish the post from Shopify admin.
Blog post drafts reference
| Output | A draft article in the chosen Shopify blog, never published live |
|---|---|
| Featured image | Optional; uploaded with the draft and adds to the cost |
| Run time | A few minutes in the background |
| Insufficient credit | The call fails before anything is queued or spent |
| Blogs | Only blogs synced with StoreSEO are listed; syncing is done in the app |
| Failed runs | The job reports the error and credits are settled automatically |
Rules GitHub Copilot follows
- The agent states the expected credit use and waits for your yes before it starts a post.
- Posts are saved as drafts only; publishing is done by you in Shopify admin.
- While the job runs, the agent reports progress and never presents the draft as finished.
In GitHub Copilot
Copy the StoreSEO config into .vscode/mcp.json with the servers key, not mcpServers, which belongs to the portable .mcp.json file. Keep storeseo_generate_blog_post on VS Code's confirmation prompt, because that prompt is the hard stop on a tool that spends credit. If you use an access token, keep it out of a committed .vscode/mcp.json, since the file is shared with your team, and use input variables instead. The same shared file means a teammate who opens the workspace gets the StoreSEO server too, so tell them it spends credit.
What runs without asking. Reads such as SEO scores, store listings, settings and Google reports run straight away. Generating a preview spends AI credit but writes nothing, and StoreSEO tells the agent to say what a generation will cost before it runs one. Applying content, optimizing images, changing settings, publishing llms.txt or agents.md, writing a product description and deleting anything wait for your yes in the conversation. That yes is an instruction the agent follows, not a lock on the server, so your client's own tool-approval setting is the hard stop.
Questions
Why does Copilot Agent mode not show the StoreSEO blog tools?
Check that your VS Code is version 1.99 or later, that the server is running in MCP: List Servers, and on Copilot Business or Enterprise that an admin has enabled MCP servers in Copilot policy. Then confirm the blog tools are selected in the tools picker.
Does Copilot publish the blog post on my store?
No. StoreSEO saves it as a draft article in the Shopify blog you chose, and you publish it from Shopify admin.
Will the agent publish a blog post on my store?
No. StoreSEO saves generated posts as drafts in the Shopify blog you choose. You review and publish them from Shopify admin.
How long does a blog post take?
A few minutes. The writing runs in the background and the agent checks the job until it is completed, failed or cancelled.
What does a blog post cost?
It spends AI Content Optimizer credit, and the amount grows with the word count and with a featured image. If you do not have enough credit the call fails before anything is spent.
Why does the agent say I have no blogs?
Only blogs synced with StoreSEO are listed. Sync your blogs in the StoreSEO app, then ask again.
Keep going
Blog post drafts with other agents
More GitHub Copilot guides