YAFL

Agent-first, end-to-end-encrypted file transfer over one MCP call or CLI command.

Free· Free to use; the schema.org offer on the homepage lists price 0 USD. No paid tiers advertised at time of review.Agents

Visit website →

Best for

Developers wiring MCP-based agents across a laptop, phone, and VPS who need a one-call way to hand encrypted files between environments without scp key juggling or cloud-drive setup.

Skip if

Teams needing large-file transfer (>100 MB), long-lived shared storage, self-hosted or open-source deployment, or contractual SLAs and audit logs for regulated workloads.

YAFL (Yet Another File Layer) is a purpose-built file-transfer layer for AI agents and the developers driving them. Its central idea is that agents living on different machines - a laptop, a VPS, a Docker container, a phone via SSH - shouldn't need scp keys, shared cloud drives, or human copy-paste to hand each other a screenshot, a PDF, a build artifact, or a zipped project directory. Instead, an agent invokes a single MCP tool call (upload_file / download_file) or the yafl CLI, and gets back a share link the receiving agent can dereference with one call of its own.

Everything is encrypted client-side with a fresh AES-256-GCM key before it leaves the machine. The decryption key travels only in the URL fragment (never transmitted to any server, and the site's Matomo config is explicitly patched to strip fragments to avoid leaking it). The app server and R2 object storage only ever see ciphertext, links auto-expire after 24 hours, and there is a 100 MB per-transfer cap. Optional password protection and one-time (burn-after-read) links are supported.

YAFL ships an MCP server (npx @yafldev/mcp), a global CLI (@yafldev/cli with put/get/status/list/delete/whoami), a Claude Desktop .mcpb bundle for one-click install, a REST API (POST /api/v1/transfers, complete, GET, DELETE), and a Claude-Code-style skill file. Login uses RFC 8628 device flow tied to a verified-email account. Common workflows: piping a screenshot from your phone into Claude Code running on a VPS, moving a build artifact between two servers that can't SSH to each other, or letting agent A drop a file for agent B to pick up asynchronously within the 24h window.

Editor's take

A refreshingly narrow tool that fills a real gap: agent-to-agent file handoff without the scp dance. The threat model is spelled out, the fragment-key discipline is respected end-to-end (even in the analytics config), and the MCP surface is small enough that agents actually use it correctly. The 100 MB cap and closed-source, hosted-only model keep it firmly in the utility bucket rather than the infrastructure bucket - but for its job, it's the tidiest option I've seen.

— The AI Tool Bible editorial team

Pros

✅ MCP-first: drops straight into Claude Desktop, Claude Code, Cursor and other MCP clients with one npx config block.
✅ True end-to-end encryption - AES-256-GCM client-side, key in the URL fragment, server only ever holds ciphertext.
✅ Async by design: 24-hour link acts as a buffer, so sender and receiver agents never need to be online at the same time.
✅ Same primitives available as CLI, MCP tools, and REST API, so scripts and agents share one mental model.
✅ Sensible security defaults: expiring links, one-time-download option, password protection, no fragments logged to analytics.
✅ Zero-setup retrieval - anyone with the link can download without an account or API key.
✅ Practical developer touches: .mcpb bundle for Claude Desktop, a shipped SKILL.md, machine-readable JSON output, honest exit codes.

Cons

⚠️ 100 MB per-transfer cap and hard 24-hour expiry rule out large datasets, model weights, or long-lived shares.
⚠️ Not open source - the server is hosted by YAFL, so security-conscious teams must trust operator claims about ciphertext-only storage.
⚠️ Free service with no stated SLA or paid tier; production reliance is at your own risk if traffic or the project changes.
⚠️ Requires Node 20+ and account creation with a verified email for uploads, which is friction for locked-down enterprise laptops.
⚠️ Losing the URL fragment means losing the file forever - there is no server-side recovery path by design.

Use cases

Agent-to-agent file handoff via MCPScreenshot from phone into Claude Code on a VPSServer-to-server file transfer without SSH key forwardingSending build artifacts out of a headless agent sessionOne-time encrypted link sharing with a colleaguePassword-protected file drop between two developersAsync file exchange between AI agents running on different schedulesPiping generated PDFs or reports from a remote agent to a local inbox