The automation your team keeps putting off, built from three weeks for a fixed fee.
Runs on your stack, handed over with a runbook and training.
“My team uses it every week.”
yannik kopp · cso / coo · loriot.io
client reference
iot core network · zurich- 01enterprise deployment readiness
- 02retrieval over internal docs
- 03routing on ambiguity
- 04human-in-the-loop approvals
what you can buy
Caelith Labs is my freelance trade name. I train your team, set up the agents, and keep them running.
In German, Spanish and English.
Six rungs, from a €199 workflow review to a monthly care plan.
self-serve workflow review
€199One workflow mapped, the fixes ranked, a roadmap back. Buy it online, no call needed first.
training and advice for your team
from €1,500Team workshop from €1,900. Team audit €1,500, one week. Fractional head of AI from €3,000 a month.
fixed-fee builds, from three weeks
from €7,000Operations sprint in 3 weeks, automation expansion in 3 to 4, agent build in 4 to 5. Optional monthly care after handover.
Founding prices, where listed, are for the first 2 clients on each offer, in exchange for a case study, a 90-second video testimonial and a LinkedIn intro. Agent operations has 1 founding slot left.
start and train: all prices€199 review, team workshop, team audit, fractional head of ai
self-serve workflow review.
One workflow that leaks time. I map it, rank the fixes, and send back a roadmap. Buy it online, no call needed first.
team workshop
- 01Half-day, remote, anywhere: €1,900
- 02Full day in person in Berlin or Madrid: €3,500, travel at cost elsewhere
- 03Founding price for the full day: €2,500
- 0450% of the fee is credited against a team audit or fractional engagement booked within 30 days
team audit
- 01Where your team uses AI today, and where it breaks
- 02What to train, what to automate, what to leave alone
- 03A written report and a 90-day roadmap, in order of effort and payoff
- 04Fixed fee, one week
- 0550% workshop credit if booked within 30 days of a workshop
fractional head of ai
- 01Two half-days a month plus async support
- 02Sets your AI priorities, tool choices and rules of use
- 03Reviews what your team builds and unblocks the people already using AI
build: all pricesoperations sprint, automation expansion, agent build, monthly care
operations sprint
- 01CRM architecture in Attio
- 02Pipelines, stages & migrations
- 03Ops dashboards & reports
- 04Team training
- 05Handover doc
automation expansion
- 01Document automation
- 02Email orchestration
- 03Workflow stitching in n8n
- 04Vendor integrations (DATEV · A3 · Holded · Idealista · BiPRO · WhatsApp Business)
- 05Quality gates
agent build
- 01Scoped task agent on Claude
- 02Guardrails & policy
- 03Evals & test set
- 04Human-in-the-loop review
- 05Ops runbook
agent operations care.
I keep what I built running: eval re-runs, runbook upkeep, fixes.
how a build runsfive phases, same order every time
discovery
Map current state. Listen for actual friction.
proposal
Fixed scope, fixed price, fixed week-count. Signed.
build
On your stack. Daily commits. Weekly walk-throughs.
verify
Real work runs through. Friction → fix list.
handover
Runbook, training, source. Owned by your team.
stack and data handlingai engine, integrations, eu infrastructure
ai engine
- 01LLM provider for production agents
- 02Anthropic listed in the /privacy sub-processor registry
- 03EU-US Data Privacy Framework (Anthropic certified)
- 04Eval suite + audit trail per build
- 05No model training on API inputs (Anthropic default)
integrations
- 01Microsoft 365 via Graph API (OAuth-scoped)
- 02Outlook inbox rules + signed envelopes
- 03Zapier , connect via public webhooks API
- 04n8n , native via webhooks REST
- 05Webhooks REST + HMAC-signed events
infrastructure
- 01Railway EU-West region (Netherlands)
- 02PostgreSQL 16 with row-level security
- 03Encrypted at rest + in transit
- 04Daily snapshots, 7-day point-in-time
questions before getting in touch
in person or remote?
Full-day workshops run in person in Berlin or Madrid; elsewhere I travel at cost. Half-day workshops, team audits, fractional work and agent builds run remote.
does the workshop fee count toward more?
50% of a workshop fee is credited against a team audit or a fractional engagement booked within 30 days of the workshop. It is the only bundle credit.
what is a fractional head of ai?
A named person who acts as your head of AI for part of the month: two half-days plus async support. They set priorities, choose tools, review what the team builds and answer questions. Three-month minimum.
which languages?
Workshops and calls run in German, Spanish or English, in the language your team works in. This website is English only for now.
work directly with me
work directly with me.
Caelith Labs is my freelance trade name. I build automation, produce explanatory videos and review workflows. I handle the work directly, from the first conversation to the agreed handover.
berlin · in person in berlin and madrid, otherwise remote
proof and a free checklist
i run two media projects with ai agents.
@vektor.fm and @huberttheinventor. Agents research, write, build and cut the films. I review the work before publication.
These are my own projects. They are not evidence of client revenue or time saved.
see how it is builtthe free 15-minute admin audit.
Find one recurring task worth examining. No call, account or email address needed.
open the free admin audit