ParloParlo

Guides

Cloud drives

Store, stream, cache, and share regular files from the Parlo filesystem.

Parlo drives are cloud-backed folders mounted next to your API connections.

bash
4 lines
1~/parlo/2  main/3  github/4  linear/

Use drives for normal files: PDFs, videos, screenshots, datasets, exported reports, generated artifacts, and agent working files.

bash
3 lines
1cp report.pdf ~/parlo/main/reports/2vim ~/parlo/main/notes/ideas.md3open ~/parlo/main/videos/demo.mp4

Streaming files

Large files are listed quickly and read in chunks. When a program asks for a byte range, Parlo fetches only that range instead of forcing a full download first, so video players, PDF viewers, and scripts get the bytes they need as they go. When Parlo detects sequential reads it prefetches ahead, which keeps playback and start-to-end scans smooth.

bash
1 line
1open ~/parlo/main/videos/demo.mp4

This is what makes videos, archives, logs, model files, and large exports usable straight from the mount.

Local cache

Parlo caches chunks locally so repeat reads skip the download.

bash
2 lines
1parlo cache status2parlo pin ~/parlo/main/videos/demo.mp4

Pinning keeps a file warm on the machine. Use it for demos, media, datasets, or anything you reuse often.

Write-through uploads

Saving or copying into a drive uploads to Parlo Cloud.

bash
2 lines
1printf "release notes\n" > ~/parlo/main/notes/release.md2cp ./dataset.csv ~/parlo/main/data/dataset.csv

Share links

Create a public link for a file:

bash
1 line
1parlo share ~/parlo/main/reports/report.pdf

Access drives via MCP

Drives are accessible through the Parlo MCP server, not just via mount or CLI. toolList returns drives alongside API connections. toolUse accepts a drive target for list, read, stat, and write operations:

json
3 lines
1{ "operation": "list", "drive": "main", "path": "/" }2{ "operation": "read", "drive": "main", "path": "/reports/notes.md" }3{ "operation": "write", "drive": "main", "path": "/reports/notes.md" }

This lets MCP-only clients (Cursor, Claude Code via MCP) browse and edit drive files the same way they access API connections.

Why drives live beside APIs

Many tasks need structured systems and raw files together. Read a ticket, generate an export, save it to a drive, share it, and let another agent or teammate continue from the same folder.

← Previous

API connections

Next →

Codebase search