Private Beta — Launching Jan 27, 2027

Deploy
AI agents.
Pay for
what runs.

SubterminalAgents is the open marketplace for production-grade AI agent applications. Download, deploy, and scale intelligent agents with transparent per-token and per-use pricing — no subscriptions, no surprises.

Browse Agents ↓ Read the Docs

subterminal — bash

$ sta install research-agent@latest

→ Resolving dependencies...

✓ research-agent v2.4.1 installed

$ sta run research-agent --task "Analyze Q4"

→ Agent initialized — model: sta-pro-v3

→ Tokens: 14,821 input / 3,204 output

→ Cost: $0.0048 — billed to workspace

✓ Report saved → ./output/q4-analysis.md

340+

Agent Apps

$0.0002

Per 1K Tokens

99.97%

Uptime SLA

Jan 27

Launch 2027

Agent Marketplace

Find the right agent
for every workflow.

🔍

Featured

DeepResearch Agent

Autonomous multi-step research. Browses, synthesises, and generates structured reports from hundreds of sources in minutes.

researchweb-browsereports

$0.006 / run

Download

</>

Hot

CodeReview Pro

Full PR review agent. Detects bugs, security vulnerabilities, performance regressions, and suggests refactors with context-aware explanations.

codesecuritygithub

$0.0014 / 1K tokens

Download

⬡

New

DataPipeline Agent

Natural-language data transformation. Connect CSVs, databases, or APIs and describe your transformation — the agent writes and executes the pipeline.

datasqletl

$0.003 / pipeline run

Download

✍

Beta

ContentForge Agent

Multi-format content agent. Blog posts, social copy, press releases, and email sequences — all brand-voice consistent using your style guide.

writingSEObrand

$0.0008 / 1K tokens

Download

🛡

Featured

SecAudit Agent

Autonomous security auditor. Scans codebases for OWASP Top 10, CVEs, secrets, and misconfigured infrastructure. Outputs prioritised remediation plans.

securitydevsecopscompliance

$0.012 / audit

Download

◉

New

KnowledgeBase Agent

Ingests docs, PDFs, and wikis to create a conversational knowledge base with citations. Continuous sync from Notion, Confluence, and Drive.

ragembeddingsknowledge

$0.0004 / query

Download

View all 340+ agents →

How It Works

Agents deployed
in four steps.

Browse & Select

Search by capability, model, or pricing model. Read reviews, inspect token budgets, and preview execution logs before committing.

Download & Configure

One command installs any agent. Configure via YAML manifest — API keys, memory backends, tool permissions, and output formats.

Run & Monitor

Execute locally, in Docker, or via managed cloud. Every token consumed, every tool called, every decision logged in real time.

Pay Per Use

No subscriptions. Billed per-token on LLM calls plus per-run fees set by the agent developer. Itemised usage visible to the cent.

Platform Capabilities

Everything agents need
to run in production.

⚡

Streaming Execution

Real-time token streaming via WebSocket. Watch agents think, decide, and act step-by-step. Built for interactive applications and debugging complex agent loops.

◈

Persistent Memory

Agents remember context across sessions using vector and episodic memory backends. Choose Redis, Pinecone, or our managed memory layer — billed per embedding stored.

⚙

Tool Registry

Browse 200+ pre-built tools — web search, code execution, file I/O, APIs, browsers — or publish your own. Fine-grained permission controls on every tool call.

◉

Usage Observatory

Per-run cost breakdowns, token attribution by step, latency traces, and model comparison dashboards. Export to Datadog, Grafana, or your own observability stack.

⌨

SDK & API Integration

First-class Python and TypeScript SDKs. Embed SubterminalAgents into your product with a few lines of code. Full OpenAPI spec available for custom integrations.

Read the SDK docs →

import subterminal from '@sta/sdk'

const agent = await subterminal.load('[email protected]')

const result = await agent.run({
  task: 'Summarize AI news this week',
  maxTokens: 8000,
  stream: true,
  onToken: (tok) => process.stdout.write(tok)
})

// cost: $0.0042 | tokens: 18,432

Transparent Pricing

Pay exactly
for what you use.

All plans include full marketplace access. LLM inference is billed at cost plus a small platform fee. Agent developers set their own per-run fees.

Starter

^$0_/month

For individuals exploring AI agents. Pay only for tokens and runs consumed.

✓ Access to 340+ agents
✓ $10 monthly token credit
✓ 5 active agent sessions
✓ Community support
✓ Usage dashboard
— Team workspaces
— Priority execution

Available at Launch

Best

Pro

^$29_/month

For power users and small teams. Reduced token rates and priority GPU queuing.

✓ Everything in Starter
✓ 20% discount on all tokens
✓ $50 monthly token credit
✓ Unlimited agent sessions
✓ Priority execution queue
✓ Up to 3 team members
✓ Advanced analytics

Available at Launch

Enterprise

Custom

For organisations with high volume, compliance requirements, or custom agent needs.

✓ Everything in Pro
✓ Volume token discounts
✓ Private agent registry
✓ SOC 2 + HIPAA available
✓ Dedicated infrastructure
✓ 99.99% SLA guarantee
✓ Dedicated success manager

Contact Sales

Model

Input / 1K tokens

Output / 1K tokens

Context

STA-Ultra

Highest capability

$0.0080

$0.0240

200K

STA-Pro

Balanced power

$0.0020

$0.0060

128K

STA-Fast

Speed optimised

$0.0002

$0.0006

32K

STA-Embed

Embeddings only

$0.0001

—

Product Roadmap

Building toward
January 27, 2027.

Our public roadmap. We ship fast and ship open — every milestone is tracked and accountable.

Q1 2026 — Completed

Foundation Done

Core platform architecture, agent runtime engine, authentication, billing infrastructure, and developer CLI tooling.

Agent Runtime v1CLI (sta)Auth & BillingToken Metering

Q2 2026 — Completed

Marketplace Alpha Done

First 50 agents published by internal team. Marketplace UI, search, reviews, and download infrastructure launched to 200 alpha testers.

Marketplace UI50 AgentsReviewsSDK v0.1

Q3 2026 — In Progress

Private Beta Active

Expanding to 2,000 beta users. Third-party agent publishing opens. Team workspaces, persistent memory, and tool registry shipping now.

3rd-party PublishingTeam WorkspacesMemory LayerTool Registry

Q4 2026 — Planned

Enterprise Preview Upcoming

SOC 2 Type II certification, HIPAA compliance, dedicated infrastructure tiers, SLA agreements, and enterprise sales motion live.

SOC 2 CertDedicated Infra99.99% SLAAudit Logs

Dec 2026 — Planned

RC1 Feature Freeze Upcoming

All v1 features locked. Security audit, load testing to 10,000 concurrent runs, documentation complete, and 340+ agents in the marketplace.

Pen TestingLoad TestingDocs Complete340+ Agents

Jan 1–26, 2027 — Planned

Soft Launch Upcoming

Invite-only expanded access. Onboard first 10,000 users. Monitor infrastructure under production load. Developer partner programme announced.

10K OnboardingPartner ProgrammePress EmbargoInfra Scaling

January 27, 2027

Public Launch Launch Day

SubterminalAgents opens to the world. Full marketplace, all pricing tiers, developer publishing, and enterprise sales live simultaneously.

Public AccessPress ReleaseProduct HuntDeveloper Blog

Launching January 27, 2027

SubterminalAgents opens publicly on January 27, 2027. Currently in private beta — developer access invitations are being issued ahead of launch.

The Team

Built by founders who
have shipped at scale.

Three engineers who have built and deployed AI infrastructure across leading technology organisations.

Raman

CEO & Co-founder

Previously led AI infrastructure at scale. Deep expertise in distributed systems, LLM inference optimisation, and developer tooling. Passionate about making agent computing accessible to every engineer.

Jayson

CTO & Co-founder

Systems engineer with a background in building high-throughput API platforms and multi-agent orchestration frameworks. Leads platform architecture, the runtime engine, and developer SDK design.

Ashley Luu

CPO & Co-founder

Product leader with a track record of launching developer-first platforms. Shapes the marketplace experience, pricing design, and the end-to-end journey from agent discovery to production deployment.

Documentation

Built for developers
who move fast.

Comprehensive guides, API reference, and SDK docs. From your first sta install to running 10,000 agents in production.

Quick Start

Get from zero to running your first AI agent in under five minutes. This guide covers installing the CLI, authenticating your workspace, and running a pre-built agent from the marketplace.

ℹ Node.js ≥ 18.0 or Python ≥ 3.10 required. API keys are provisioned automatically on account creation.

1. Install the CLI

$ npm install -g @subterminal/sta

# or: pip install subterminal-agents

2. Authenticate

$ sta auth login

✓ Logged in as [email protected]

3. Install and run an agent

$ sta install research-agent

$ sta run research-agent --task "your task"

Full documentation →

DeployAI agents.Pay forwhat runs.

Find the right agentfor every workflow.

Agents deployedin four steps.

Everything agents needto run in production.

Pay exactlyfor what you use.

Building towardJanuary 27, 2027.

Built by founders whohave shipped at scale.

Built for developerswho move fast.

Agents are thenew applications.

Deploy
AI agents.
Pay for
what runs.

Find the right agent
for every workflow.

Agents deployed
in four steps.

Everything agents need
to run in production.

Pay exactly
for what you use.

Building toward
January 27, 2027.

Built by founders who
have shipped at scale.

Built for developers
who move fast.

Agents are the
new applications.