FastSend AI · MCP

Turn file delivery into an action your AI assistant can perform

Ask your AI to send a document, create a protected file-collection link, check whether a client opened a transfer, see the exact download time or wait for a reply. FastSend performs the action, keeps an audit trail and enforces limits on the server.

“Send this PDF to the client and tell me when they download it.”
“Create a link where accounting can upload the documents and wait for a reply.”
“Show what happened to yesterday’s transfer: sent → opened → downloaded → replied.”

Send files

One MCP tool runs the existing protected FastSend pipeline: single PUT or multipart, confirmation and finalize.

Collect files

The AI creates a document-intake link. Incoming submissions are visible through the same MCP and event layer.

Precise status

Events are separate from webhooks: sent, opened, downloaded, reply received, or a new submission arriving in a collection.

MCP Free

0 ₽
for evaluation and smaller tasks
  • Up to 3 GB per file
  • 10 GB of AI traffic per month
  • Up to 10 transfers per hour
  • Up to 5 recipients per transfer / 25 unique recipients per 24 h
  • File collection, status tracking and replies
FastSend AI

FastSend AI

250 ₽ / month
first month — 119 RUB
  • Up to 10 GB per file
  • 50 GB of AI traffic per month
  • Up to 60 transfers per hour
  • Up to 25 recipients per transfer / 250 unique recipients per 24 h
  • Extended retention up to 30 days
Confirm email and connect FastSend AI The 119 RUB first-month price and eligibility are verified on the server after email confirmation.

Connect your AI client

Use the cloud connection for online AI tools. Use the local connection when the assistant needs to read files on your computer.

https://fast-send.ru/mcp

Codex · cloud

No Python installation or API key is needed. Small generated files can use text/base64 (up to 1 MiB). Large files upload directly to storage using MCP tools or the account-bound prepare_upload page. Free defaults to 3 GiB per file; quote_upload returns actual limits and prices. The user pays for extra volume or FastSend AI in a browser; verified purchases activate automatically on the same account.

codex mcp add fastsend --url https://fast-send.ru/mcp
codex mcp login fastsend
codex mcp list

ChatGPT

  1. Enable Developer mode in ChatGPT settings if your account allows it; open Plugins and add a connection.
  2. Enter the server URL: https://fast-send.ru/mcp
  3. Confirm your email and approve the requested permissions on FastSend.

Claude

  1. Open Customize → Connectors → Add custom connector.
  2. Enter the URL. For OAuth registration choose Register automatically: https://fast-send.ru/mcp
  3. Confirm your email and approve the requested permissions on FastSend.

Cursor

Add this remote server in Cursor MCP settings or your mcp.json file:

{"mcpServers":{"fastsend":{"url":"https://fast-send.ru/mcp"}}}

Local files: Codex and desktop clients

Extract the archive and run Install-Windows.cmd on Windows. The installer asks for the key using hidden input and creates mcp.json for LM Studio, Claude Desktop and Cursor, plus codex.toml. Merge the fastsend entry into your client settings and restart it. Python 3.11+ is required. Credentials are stored separately with restricted permissions. No background polling is used.

Download the local MCP client

Windows

cd fastsend-mcp
Install-Windows.cmd

Setup generates mcp.json and codex.toml. For Codex, merge codex.toml into your configuration. Example using a separate credentials file:

# ~/.codex/config.toml
[mcp_servers.fastsend]
command = 'C:\path\to\fastsend-mcp\.venv\Scripts\python.exe'
args = ["-m", "fastsend_mcp.server", "--config", "C:/Users/YOU/.fastsend-mcp/credentials.json"]
# Local files: Codex and desktop clients

macOS / Linux

cd fastsend-mcp
python3 -m venv .venv
.venv/bin/python -m pip install .
.venv/bin/python -m fastsend_mcp.configure

Setup generates mcp.json and codex.toml. For Codex, merge codex.toml into your configuration. Example using a separate credentials file:

# ~/.codex/config.toml
[mcp_servers.fastsend]
command = '/absolute/path/to/fastsend-mcp/.venv/bin/python'
args = ["-m", "fastsend_mcp.server", "--config", "/home/YOU/.fastsend-mcp/credentials.json"]
# Local files: Codex and desktop clients