Based in Lahore
Open book · Updated 2026-09-23
Prices, method and proof in public
→Open book

The open book. Everything we publish, in one place.

Our prices, our method, our proof policy, our evals and our mistakes. What's published is here. What isn't yet says so, with a date where we have one. Trust is earned in the open.

Last updated
EntryStatusWhere
Prices and termsPublishedPrices
MethodPublished, with a one-page PDFThe Deploy Ladder
Proof policyPublishedProof policy
Eval scorecardsNone published yetEvals
Post-mortemsNone published yet. The first lands by 26 Dec 2026.Post-mortems
Studio factsPublishedFacts
Who we're not forPublishedNot for
ChangelogOne entryChangelog
01Prices

Two engagements. Both fixed scope, fixed price.

Engagement
First Agent
$8K — 15K
4 — 6 weeks
For

One workflow, one agent (or a small team of agents) live in production.

Engagement
Production System
$15K — 30K
6 — 10 weeks
For

A multi-agent system with orchestration, evals, observability and a client-facing dashboard.

Terms for both.

  • Fixed scope, fixed price. No hourly billing.
  • SOW signed before any work.
  • 50% at SOW signing (non-refundable), 50% at production handoff.
  • IP transfers on final payment.
  • Two iteration rounds.
  • 30-day support after handoff.
  • Change requests get their own SOW.
  • The 30-min scope call is free and scopes the work. No paid discovery phase before signing.
02Method

The Deploy Ladder. Public, down to what "done" means.

  • PhasesMap · Build · Wire · Ship
  • RungsInternal preview, staging on main, production behind flags
  • First FlipUsually within ten working days of kickoff
  • CadenceDaily written update, one weekly sync, a Friday demo
  • Handoff boxRunbook, eval report, architecture diagram, Loom, 30-day support calendar
Read the method One-page PDF
03Proof policy

Real work. Numbers only with sign-off.

The policy

Verticals, durations, stacks and architecture are real. Outcome numbers are published only with client sign-off, and none are signed off yet.

See the five shipped systems
  • Each case study names what we measured, and says the number goes up when the client signs off.
  • Client names are withheld by default.
  • No testimonials until real ones exist.
  • No client logos. No "trusted by".
04Evals and post-mortems

Our evals and our mistakes. Nothing here yet, and we'll say so.

Eval scorecards

Scorecards from shipped systems.

Every agent we ship has an eval that proves it works. The scorecards go public one system at a time, with the client's sign-off.

None published yet. The first goes up with client sign-off.

Post-mortems

What broke, and what we changed.

When something we ship goes wrong, we write it up here: what happened, why, and what changed in the method because of it.

None published yet. The first lands by 26 Dec 2026.

05Facts

The studio, in facts. No rounding up.

Since2023. Started as contract work; registered as Codantrix Labs in 2025.
RegisteredWith the SECP in Pakistan, 2025.
Systems shippedFive, across five verticals. Property operations and field service lead.
CapacityThree engagements at once, maximum.
LocationLahore, Pakistan (UTC+5). Working with US and EU teams, async-first.
Engineering leadHassan Ali Mehdi. Seven years coding, three on production agentic systems. Founder bio
SizeSmall on purpose.
StackPython · LangGraph · pgvector · Claude · OpenAI · Cloudflare Workers · D1. Per engagement: Next.js, AWS, Milvus.
Contactcontact@codantrix.com
06Who we're not for

Who we're not for. Said early, so nobody wastes a call.

The reasons, on the studio page

07Changelog

What changed in the book. Newest first.

  1. Opened the book.

The book is open. Bring your workflow.

30-min scope call, free. By the end you'll know what we'd build, in what order, what it costs, and whether we're the right fit.

Book a 30-min scope contact@codantrix.com