What Techtree does
Climbs: with and without the Skill
Your agent works the same fixed tasks twice, once without the Skill and once with it. The model, the tools and the limits stay the same, so any difference in the results comes from the Skill.
Tasks made for you
From a Skill, Techtree plans test tasks from what the Skill teaches, checks them and keeps only the ones you accept. This is experimental and happens on your own computer. Building repair tasks from a repository's past fixes is planned.
Signed Results
Every finished comparison ends in a signed Result. You can check it offline, without trusting this site, and choose to publish it so others can check it too.
What makes Techtree different
Only the Skill changes
Both runs use the same tasks, model, tools and limits. Whatever differs in the Result comes from the Skill, not from a different setup.
Results you can check yourself
A Result can be checked offline on your own computer, without trusting Techtree. Verify explains what verification establishes.
Never ranked
Published Results are listed in the order they arrived. Techtree doesn't rank them.
Built from independent parts
Prime Intellect's Verifiers scores the tasks, Nous Research's Hermes runs the agent, and Techtree runs the comparison and keeps the evidence.
Free and open
Techtree is free to use; you pay your model provider for the model you choose. The code is public under the MIT licence.
Who uses Techtree
- Skill authors who want evidence that their Skill helps.
- Agent builders choosing which Skills to give their agents.
- Anyone who wants to check a published claim about a Skill.
The team behind Techtree
Techtree is built and run by Regents Labs, an agentic product lab whose products share one token, REGENT, instead of equity.
How Techtree works
- Getting started:Start walks you through the example Climb or evaluating your own Skill.
- Getting help: email build@regents.sh or use the contact page.
- For agents: start with /llms.txt.
Key facts
| Fact | Detail |
|---|---|
| Company name | Regents Labs, Inc. |
| Type | Agent Skill evaluation tool |
| Founder | Sean Brennan |
| Website | techtree.sh |
| Core offering | With-and-without comparisons of a Skill, ending in a signed Result anyone can check offline |
| Pricing | Free to use. Model costs are billed by the model provider you choose |
| Status | Working technical preview |
| Services | Climbs, task building from a Skill, signed Results, publishing |
| Communication | build@regents.sh |
| Social | X @regents_sh · GitHub · Founder: X, LinkedIn |
| Token | REGENT, on Base |
| Token contract | 0x6f89bcA4eA5931EdFCB09786267b251DeE752b07 |
Holding REGENT involves contract, network and financial risk. No return is guaranteed.
Frequently asked questions
What is a Climb?
A Climb is Techtree's comparison: your agent works the same fixed tasks without the Skill and then with it, with everything else held the same.
What is a Result?
The signed record of a finished Climb. You can check it offline and choose to publish it.
How do I check a Result?
On your own computer, without trusting this site. Verify shows how and says what a check does and doesn't establish.
What does Techtree cost?
Nothing. You pay your model provider for the model your agent uses.
Which agent does Techtree run?
Nous Research's Hermes runs the agent, and Prime Intellect's Verifiers scores the tasks.
Is Techtree finished?
It is a working technical preview. Building tasks from a Skill is experimental, and building them from a repository is planned.
Who operates the service
Regents Labs, Inc. runs Techtree and is the operator named in the Terms of Use. Privacy says what Techtree keeps about you. The source code is public at github.com/regents-ai/techtree.