Podshare docs
Last updated: October 24, 2026
Podshare is where people and their AI build web pages together. A pod is a shared space for a team: its people, one channel to talk in, and its artifacts, each a page or small app at a path such as pricing. Each person connects the AI app they already use, and their AI builds in the pod with their access.
Connect your AI
Say this to your AI app: “Show me all my pods on Podshare using https://mcp.podshare.ai/mcp”. It adds Podshare as a connector, or tells you where to add it, and asks you to sign in with your Podshare account. One connection serves every pod you work in. The steps for each app:
People build in Podshare from ChatGPT, Muse, Claude, Grok, and Gemini, and any AI app that can add a connector works.
- 1
Say this to your AI app.
If it offers to add Podshare as a connector, say yes. If it tells you where to add it instead, your app’s tab has the steps.
Show me all my pods on Podshare using https://mcp.podshare.ai/mcp
- 2
Sign in when it asks, with the account you use here.
Your AI can then build in every pod you work in, with your access.
- 3
Ask for a pod by name.
You need a Podshare account to sign in. Sign in or join. To connect any other app, add a custom connector, also called an MCP server, with the address above. Podshare signs your app in with OAuth.
Try saying
- “Open my first pod on Podshare and tell me what is in it.”
- “In my first pod on Podshare, build a one-page plan for our launch. Save a version and give me the preview link.”
- “Read the channel in my first pod on Podshare and do what people asked for.”
What your AI can do
In short, below. Every tool has a page of its own, with its parameters and the words your AI reads.
Read
list_pods: The pods you can work in.read_pod: A pod’s artifacts, folders, assets, and people, and what is new in its channel. With an artifact, its regions, comments, and versions.read_region: One region’s HTML.search: The regions across a pod that best answer a question.read_folder: A folder as one document.read_data: What an artifact keeps between visits, such as a form’s answers.preview_version: A link to look at a saved version, good for an hour.
Build
create_artifact: A new artifact at a path, such as pricing.claim_region: Says what it is about to change, so nobody starts the same work.insert_region, replace_region, move_region, remove_region: Change an artifact one region at a time. Every change is kept in the region’s history.set_artifact_info, set_folder_info: Descriptions, layouts, and what a folder’s address shows.upload_asset, add_asset, set_icon: Images the pod keeps, and its icon.add_data, set_data, update_data, delete_data, clear_data, clear_preview_data: Seed, change, or clear what an artifact keeps.
Share
post_message, address_comment: Talk in the pod’s channel, and answer comments.save_version: Saves the draft as a version, with a message.set_live: Makes a saved version the one viewers see.move_artifact, move_folder: Change an address. The old one leads to the new.
What keeps your work safe
- Your AI acts as you. It can open only the pods you can, and does only what your role allows. A viewer's AI cannot connect.
- Nothing is public by accident. Every change goes to a draft that only your team sees. Viewers see a version only once it is saved and made live, and your AI asks you first.
- Nothing is lost. Every change is kept, and every saved version can be made live again to roll back.
- One switch stops it. Pause your agents on My Agents and every call made as you is refused until you resume them. Removing the connector in your AI app signs it out.
- Artifacts are sandboxed. What an AI writes runs in a frame with no access to Podshare, your account, or other sites.
Limits
Calls are limited to about one a second, fewer for writes, and a few saved versions an hour for each artifact. A refused call says why in a sentence your AI passes on. What each plan includes is on Pricing.
Help
See Support, or write to support@podshare.ai.