Skip to content

Prompt Templates

These prompts work in Claude Desktop, Cursor, Moraya, and any MCP-compatible client. Copy, adapt to your context, and run.

Most prompts assume HTTP OAuth MCP is connected — see Claude Desktop setup if not.

1. Publish current README to Picora

Upload the README.md I'm currently editing as a public Picora Markdown
document. Use the title from the H1 heading. Tag it with "release" and
the current year. Return the public URL when done.

What it does:

  • Reads your current file
  • Calls upload_doc with isPublic: true
  • Embeds inline images get auto-rewritten
  • Returns the share URL

2. Bulk upload a folder of images

Upload all PNG and JPG files in the current `screenshots/` folder to
Picora. Make them public. Tag each with the folder name. Show me a
table of (filename, Picora URL, size) when done.

What it does:

  • Iterates over local files (Cursor / agent mode handles this best)
  • Calls upload_image for each
  • Renders results table

For large batches (>20 files), the AI may suggest pausing to confirm midway through — answer “continue”.

3. Rotate / replace an existing image with a new file

The image at https://media.picora.me/xK9mR2pQ7vB.jpg should be replaced
with `assets/cover-v2.png`. Upload the new file, then update all `.md`
files in `docs/` to point to the new URL.

What it does:

  • Calls upload_image with the new file
  • Greps for the old URL in .md files
  • Edits each file to use the new URL
  • Shows you the diff for approval

The old image stays on Picora unless you ask to delete it.

4. Audit my image library by size

List my Picora images sorted by size descending. Show the top 20.
Group anything over 5 MB and summarize: how much storage would I free
up if I deleted them?

What it does:

  • Calls list_media with type=image
  • Sorts client-side
  • Renders a markdown table
  • Computes the savings

This is great for finding accidental large uploads.

5. Clean up old test data

Find Picora resources I uploaded with tag "test" or "wip". List them
first. Then delete all of them — but show me the dry-run output before
actually deleting.

What it does:

  • Calls list_media with tag=test and tag=wip
  • Calls delete_media with dryRun: true, shows what would happen
  • Asks for your confirmation
  • Calls delete_media again with dryRun: false

Always require dry-run for destructive operations — this prompt does it correctly.

6. Track a long-running upload session

For the next 10 minutes, every time I drag an image in chat, upload it
to Picora as a private resource. After uploading, give me the URL but
don't open the image. At the end, summarize: how many files, total
size, total storage used after?

What it does:

  • Stays in “watch” mode for the duration
  • Calls upload_image for each drag
  • Tracks aggregates
  • Calls usage:read to compare totals

Useful for batch documentation work.

7. Convert legacy markdown blog posts to Picora-hosted

For each `.md` file in `posts/`:
1. Find all data:image/...;base64 references
2. Upload it to Picora as a public Markdown document with auto image
   rewriting (rewriteImages: true)
3. Replace the original file's content with a single line:
   "Published to Picora: <URL>"


Confirm each conversion before moving to the next file.

What it does:

  • Iterates over .md files
  • Calls upload_doc with each
  • Inline base64 images become standalone Picora image URLs (auto)
  • Replaces the local file with a stub
  • Pauses between files for your approval

Result: your local repo becomes a thin index pointing to Picora-hosted content.

8. Search my docs for keywords

Find all my Picora Markdown documents with the tag "tutorial". Show
their titles, word counts, and last updated dates.

What it does:

  • Calls list_docs with tag=tutorial
  • Renders a table

For full-text search across document content, use the dashboard’s search box. (Full-text via API is on v0.20+ roadmap.)

9. Check quota & generate report

Tell me my current usage across all Picora resources: image storage,
video storage, audio storage, document count. Compare to my plan
limits. Highlight anything over 80%.

What it does:

  • Calls get_usage
  • Renders a status table with red/yellow/green indicators

Run this monthly as a billing-cycle check.

10. Migration: pull all images, delete from Picora

Download all my Picora images to a local folder `picora-export/`. After
each successful download, delete the image from Picora. Skip any that
fail to download (don't delete them).


Show progress periodically (every 10 images).

What it does:

  • Calls list_media with pagination
  • For each: HTTP fetches the image, saves locally
  • Calls delete_image only on successful downloads
  • Reports failures

This is the “self-service migration” workflow if you ever want to leave Picora.

Adapting templates

These prompts use specific values (folder names, tags, etc.). Replace them with your own. Keep the structure:

  1. Specify the resource type (images / docs / etc.)
  2. State the action (upload / list / delete)
  3. Add filters / constraints (size, tags, date)
  4. Request output format (table, summary, just the URL)
  5. For destructive operations: explicit dry-run / confirmation step

If a prompt isn’t working as expected, ask your AI: “explain what tools you’d call for this prompt” — it’ll show its plan, and you can refine.