A person starting the workday at a sunlit desk with a laptop and steaming coffee

Treg: A Pay-Per-Call Tool Catalog for AI Agents

Treg is a catalog of paid tools for AI agents, plus a shared vault for a team’s own tools. With one base URL and one token, an agent can search more than 4,100 tools from 127 providers by the job it needs done, such as SEO and keyword data, backlinks, social media profiles and trends, people and company enrichment, ad libraries, web scraping and image, video and voice generation. It reads the price, makes the call and pays from a prepaid balance, with no account at the provider. Teams can also register their own API keys, logins and command-line tools so every teammate’s agent can use them without the credentials leaving the server. Treg calls itself “OpenRouter, but for agent tools instead of models.”

The quick answer: Treg is a well-designed, transparent option for giving agents broad access to paid data and generation tools without buying a subscription for each one. It stands out for pricing at providers’ own rates with no markup, for publishing measured reliability so agents can pick providers on evidence, and for letting teams share their own keys safely. It’s new, built by the team behind Superdesign, and its code is available to self-host. Everything below comes from Treg’s site, pricing, developer summary and GitHub repository.

Treg at a glance

Checked against Treg’s site, pricing page, developer summary and GitHub repository on November 13, 2026.

Treg
What it isA pay-per-call tool catalog and credential vault for AI agents
Catalog4,100+ endpoints from 127 providers, per Treg
Your own toolsAPI keys, OAuth connections, vendor CLIs and skills, shared across a team
AccessMCP server, CLI with agent skills, HTTP API
PricingProvider’s own rate, no markup; $1 free to start
CodeGitHub repo (about 7,900 stars), Apache 2.0 with a hosted-service restriction
CompanySuperdesign Dev Pty Ltd

What Treg does

Treg’s developer summary explains why it exists: “the tools real work needs sit behind subscriptions nobody buys for a single run, behind signup walls, or behind no public API at all.” Its README names examples, such as SEO, company data and sales-intelligence tools that each cost tens to hundreds of dollars a month. Treg holds those accounts and bills a fraction of a cent per call for most lookups.

The pitch to agents is “Ask for the job, not the vendor.” An agent searches the catalog for what it wants to do, for example backlinks for a domain, reads each matching tool’s price and reliability, and calls one. For common jobs like scraping a page, web search, finding an email or pulling a social profile, Treg offers routed tools that pick a provider, fall back to another when one fails, and bill only the answer that comes back.

Picking providers on evidence

A notable feature is that Treg measures the providers it sells. Its homepage notes that “Nineteen providers sell backlinks, nine sell email lookup,” and for each one Treg shows what it charges and what it has measured: success rate, speed and the last time it answered. Its guidance to agents is to choose by matching inputs first, then measured reliability, then price.

Treg’s pricing page also publishes receipts from real runs. In one, building a verified lead list from 50 companies cost $2.33 in total, or $0.12 per deliverable lead; in another, pulling search volume and trends for 50 keywords cost $0.11.

Your team’s own tools

The second half of Treg is a credential vault. A teammate can register a paid API account, an OAuth connection, a vendor CLI like Stripe’s or GitHub’s, or a skill file, and every agent on the team can call it while the credential stays on Treg’s server and every call is audited. Treg says “The team’s own key always wins over treg’s and is never billed.” A scan command shows which local keys and skills could be shared, and nothing is uploaded without approval.

This makes Treg useful beyond the catalog: a team can give agents access to its own tools without copying secrets onto every laptop.

How to connect

Treg works in three settings, which its documentation is careful to distinguish:

  • Hosted assistants like ChatGPT and Claude.ai add Treg’s MCP server as a custom connector and approve access, choosing which team pays.
  • Local coding agents like Claude Code, Cursor and Codex install a CLI that also adds Treg’s skill to every detected agent, then sign in through a browser.
  • Sandboxed agents in locked-down cloud containers may need an administrator to allow Treg’s address.

Over MCP, the core steps are searching the catalog, getting a tool’s details and calling it, with extra tools for checking the balance, listing your own tools and requesting tools the catalog lacks.

Pricing

Treg’s pricing is a prepaid balance “metered per call at the provider’s own rate, with no markup,” and Treg says the no-markup promise is in its terms. The first $1.00 is free. Calls on your own keys, tools and connected accounts are never metered. Some providers bill only for successful results, so a miss costs nothing; others bill every call. When the balance runs out, metered calls stop with a clear error, while your own-key calls keep working. There are no seats.

The code and the company

Treg was built for the Superdesign team, according to its README, and its GitHub repository, created in July 2026, has about 7,900 stars. The license is Apache 2.0 with additional terms from Superdesign Dev Pty Ltd, including a restriction on offering it as a hosted service, so companies can self-host it internally but shouldn’t plan to resell it as a competing service. We didn’t find independent reviews or announced funding.

How Treg compares

Treg is part of a new group of tool gateways that each call themselves an OpenRouter for tools, alongside Monid and ToolRouter. Monid emphasizes a huge data catalog including Chinese social platforms and focuses on reading data; Treg adds a team credential vault, measured provider reliability and published cost receipts, and includes some actions such as publishing Instagram reels. These gateways differ from action platforms like Composio, which connect agents to everyday business apps, and they resell individual providers like Exa and Firecrawl.

If what you need is an assistant that works across your own apps rather than a catalog for an agent you’re building, Carly connects to thousands of apps, so from ChatGPT, Claude or a text message it can sort your email, manage your calendar and keep your CRM up to date.

Who Treg suits

Treg is a good fit if:

  • Your agents need occasional access to many paid data and generation tools.
  • You want transparent pricing at providers’ rates, with no markup.
  • Your team wants to share API keys and CLIs with agents without spreading secrets around.
  • You want agents to choose providers by measured reliability and price.

Look elsewhere if:

  • You need deep actions inside business apps like Gmail, Salesforce or Slack.
  • You want to offer the software as your own hosted service.
  • You want a finished assistant rather than infrastructure.

FAQ

What is Treg?

Treg is a catalog of more than 4,100 paid tools from 127 providers that AI agents can search and call with one token, paying per call from a prepaid balance. It also lets teams share their own API keys and tools with their agents securely.

How much does Treg cost?

Treg charges each provider’s own rate per call with no markup, from a prepaid balance. The first $1.00 is free, and calls using your own keys are never metered.

Does Treg work with ChatGPT and Claude?

Yes. Treg’s MCP server can be added as a custom connector in ChatGPT and Claude.ai, and its CLI and skill work with Claude Code, Cursor and Codex.

Is Treg open source?

Treg’s code is on GitHub under the Apache 2.0 license with additional terms, including a restriction on running it as a competing hosted service. Teams can self-host it.

Who makes Treg?

Treg is made by Superdesign Dev Pty Ltd, the company behind Superdesign, and was originally built for the Superdesign team.

Ready to automate your busywork?

Carly schedules, researches, and briefs you—so you can focus on what matters.

See what people say

"Before Carly, I relied on a Calendly link, but the whole process felt impersonal and not very professional. Carly changed that by handling all the back-and-forth, so I'm no longer stuck in endless email threads trying to line up schedules.

Now Carly reaches out to candidates, shares my real-time availability, lets them pick a slot, then sends a Zoom link and drops it straight into my calendar. She sends reminders to both of us before each call, which has significantly reduced no-shows and last-minute confusion.

On top of scheduling, Carly acts like a full executive assistant, sending me my schedule the night before so I can prepare for each call. It reminds me of the old x.ai assistant, but Carly is noticeably smarter, faster, and better suited to my healthcare recruitment business."

Gus Ibrahim, Founder & Director, IHR