Pricing & billing
Why bring my own LLM key?
Because we shouldn't be reselling OpenAI tokens with a markup. You pay your model provider directly — we charge for the platform that makes the agents useful. Your model bill is your model bill.
Can I cancel anytime?
Yes. Cancellation is one click in the Stripe Customer Portal from your dashboard. You keep access for the remainder of the billing month. No prorated refunds, but no lock-in either.
Do you have a free trial?
Yes — 7-day free trial. Credit card required at signup. You can cancel before the trial ends and you won't be charged.
What if I hit my monthly workbench quota?
Forge is capped at 50 builds, Studio at 20 video projects, Coherence at 1,000 calls. If you bump the ceiling we'll talk — write us and we'll figure out the right shape. We aren't going to ambush you with overages.
OSS
What's in the OSS version?
The full agent framework — 220 agents, 380 tools across 35+ categories (growing weekly), the entire runtime including OSS Swarm orchestration. It runs on your hardware (Jetson, laptop, server). What it doesn't have: the hosted dashboard, Coding Forge / Video Studio / Coherence Cloud (which require our backend), or the Nexus Hub tunnel.
Is the OSS really free forever?
Yes. MIT license. Fork it, ship it inside your product, modify it however you want. No clawbacks, no commercial-use cliff. The membership is for the hosted layer + the cloud workbenches + what we ship to ourselves.
Can I pair the OSS install with my membership dashboard?
Yes — and you should. Generate a REPRYNTT_API_KEY in the dashboard, drop it in your OSS install, and the dashboard becomes a control surface for the agents running on YOUR hardware. The Nexus Hub then tunnels your local Nexus into the cloud dashboard.
Providers & cost
What LLM providers do you support?
OpenAI, Anthropic, xAI, Google Gemini, NVIDIA NIM, or local (Ollama, vLLM, anything OpenAI-compatible). Connect one or several — the platform routes intelligently between them per task.
How much will my actual LLM spend be?
Depends on what you run. Our burn tests on JARVIS at 40 RPM came in at ~$6.24/hr on Grok 4.3 and ~$10.60/hr on Claude Opus 4.6 (worst case, top-tier model as producer). Hybrid routing — cheap workhorse + premium orchestrator — brings the same caliber of work down to ~$1–1.50/hr. Cheap-tier-only runs at pennies per hour.
Can I cap LLM spend?
Yes. Set a per-provider monthly budget in Settings → LLM Providers. We halt routing to that provider when the cap is hit. No surprise invoices.
Enterprise & integration
What about enterprise / on-prem?
Email hello@ai158z.com. We'll work out a fit — on-prem deploys, custom integrations, SLA, dedicated support, air-gapped install options.
Can I integrate this into my product?
Yes. OSS is MIT-licensed. The hosted platform exposes a REST API at api.repryntt.com — Bearer token auth, all 220 agents available via /v1/agents/run, all workbenches scriptable. See /docs/api for the surface.
Do you support webhooks / inbound tasks?
Yes — /v1/inbox is the universal inbound channel. Any external system POSTs a task; the worker picks it up, runs the agent, and delivers results via callback URL or polling. Native Slack adapter included; webhooks for Zapier / n8n / etc. work out of the box.
Still have questions? hello@ai158z.com — direct reply, usually within a business day.