touchNspace

Create your account

Get a key, try the tools in the Sandbox and connect your AI agents in minutes.

Google GitHub
Or

Already have an account? Sign in

What is it

Outbound marketing tools for AI agents

touchNspace is a cloud space of tools for outbound marketing, a bit like a cloud platform but for outreach. The tools score and improve messages and rank drafts, prospects and segments.

Behind them is our own expert model, not a general chatbot. Two things follow: the same input always gets the same score, and every score explains why, with fixes your agent can apply to improve the message automatically.

The tools are built for AI agents. Connect yours over MCP, give it our SKILL.md, or paste ready-made instructions into its prompt. This website is for people: it shows what the tools do, lets you try them and connects your agents.
How it works

Three steps from sign-up to your agent

  1. 1
    Try the toolsCreate an account and run any tool in the Sandbox with sample data already filled in.
  2. 2
    Connect your agentsThree ways, pick what your agent supports: add our MCP server (Claude, Cursor, VS Code and any MCP client), drop our SKILL.md into its skills, or paste a tool’s copy-ready instructions into its prompt. Each agent gets its own access key.
  3. 3
    Follow the usageSee which agent called which tool, and how much of your allowance is left.
Why trust the scores

An expert model, not a guess

Ask a general AI to rate a message twice and you can get two different answers. Our tools run on our own expert model, so they behave like a measuring instrument:

  1. 1
    Stable, comparable scoresThe same message, recipient and offer always get the same 0–100 score. Scores sit on one scale, so you can compare drafts, contacts and campaigns, set thresholds such as “send only above 70”, and track progress over time.
  2. 2
    It explains why, the same way every timeEach score lists what raised and what lowered it, in points, plus fixes ranked by how much each would add. The reasons are as repeatable as the score, so your agent can improve the message automatically: apply the fixes, re-score, and know the gain came from its edit, until the message passes your bar.

Before a message goes out the model also flags hard problems: an unfilled personalisation token, the wrong language, a recipient outside the buyer group, no clear ask, or a message that looks like a scam. Results name the model version they came from.
Free tier

Start for free

1,000tool calls every month, included with every account

One tool call is one request from an agent or from the Sandbox. The count starts again each month, and you can follow it on the Usage page.