Pay for POMs, hours, or tokens.
Three ways to work together — a half-hour POM for one focused task, an always-on managed agent, or API access at cost-plus-£1. No retainers, no minimums.
One POM is half an hour of focused, AI-augmented build.
Fixed price, fixed scope, fixed time. Bring a problem, leave with progress. The first POM is on the house.
One POM call. Map the problem, scope the work, decide if we're a fit.
Book a free callTwo hours of paired, AI-augmented build. One small thing shipped or unblocked.
Book 5 POMsHalf-day sprint. A feature, a prototype, or a research dossier in your hands.
Book 10 POMsFull day. A complete deliverable — production-grade app, full pipeline, end-to-end.
Book 20 POMsfrom £1 / hour.
Always-on, self-driving agents that run your pipelines, generate your daily content, watch your inbox, and ship to your sites — billed by the hour, not the seat.
- Self-driving, audited, multi-model
- Pay only for hours the agent ran
- Full ownership of the pipeline + outputs
- Swap models, swap providers, anytime
PAD + LAD + IAD running every morning at 06:00, posting to your CMS by 07:00. Typical runtime ≈ 1.5 hours/day across all three.
Cost + $1/M tokens.
Pass-through API access — built on claude.com/pricing#api and equivalent OpenAI / OpenRouter rates, plus $1 per million tokens to cover routing, observability, and audit. One bill, every model.
| Family | Model | Input / 1M | Output / 1M | Notes |
|---|---|---|---|---|
| Claude | Opus 4 | $16.00 | $76.00 | Frontier reasoning · coding · agents |
| Claude | Sonnet 4 | $4.00 | $16.00 | Balanced · production default |
| Claude | Haiku 4 | $2.00 | $6.00 | Fast · cheap · long-context |
| OpenAI | GPT-5 | $11.00 | $41.00 | OpenAI flagship |
| OpenRouter | Mixed routing | — | — | Pass-through cost + £1 markup |
Claude rates pulled from claude.com/pricing#api at time of publication. Markup is $1/M tokens on both input and output.