The FORGE method

Evidence over guesswork,
one gated step at a time

FORGE is how IPGuru turns a raw idea into a filing-ready invention. Five phases, Focus, Observe, Refine, Grade, Execute, with a gate between each. Every phase names real work, produces real analysis, and ends in a clear decision before the next begins. You never build on a foundation you have not tested.

Why we work this way

Innovation is a series of good decisions

Most ideas fail not because they are bad, but because time and money are spent before anyone checks whether they should be. FORGE front-loads the cheap, decisive work and gates the expensive work behind it.

savings

Validate before you invest

Focus and Observe are free. You learn whether an idea is worth pursuing before you spend anything structural.

checklist

Name the real work

Each phase is an activity you actually do, capture, research, build, test, deliver, not a vague stage.

door_front

Gate every step

Between phases sits a gate: a short checklist and a decision. You advance only when the evidence supports it.

sync

Iterate and pivot

Weak results loop back with specific direction, and a PIVOT lets you change course without losing your work.

The journey

Five phases, four gates

Follow one invention through. Each phase shows the analysis it produces and the crew on it; each gate shows the decision that unlocks the next.

center_focus_strong

Focus

F, capture the idea
Free

What is analysed

  • arrow_right Your idea in your own words, turned into project attributes
  • arrow_right The problem it solves and who it is for
  • arrow_right A first concept image to react to, before any writing

The crew on it

Nova, guides Kira, first images
door_frontGate to Observe. The disclosure questions are answered and you have reacted to a first image. Decision: is the idea clear enough to research?
travel_explore

Observe

O, research and validate
Free

What is analysed

  • arrow_right The state of the market and who the competitors are
  • arrow_right An early patentability read: is it new, novel, non-obvious, feasible
  • arrow_right Viability, can it actually be built
  • arrow_right A preliminary web prior art search

The crew on it

Cora, research Vega, pressure test Sean
workspace_premiumGate to Refine, the upgrade gate. The gate report is in: recommendations, red flags, the hard gates of novelty and non-obviousness, the soft gates of feasibility and cost, and any overrides you logged. Decision: is this worth building? Upgrading the project here unlocks the paid phases.
handyman

Refine

R, build the deliverables
Upgrade

What is analysed

  • arrow_right The imaging loop: concept images first, the physical description derived from the ones you accept
  • arrow_right The application draft, the claims, and the technical diagrams
  • arrow_right Strengthening variations and alternative embodiments

The crew on it

Kira, imaging Remy, drafting Sean Iris, formatting
door_frontGate to Grade. Visuals accepted, the physical description confirmed, the draft and claims in hand. Decision: is it ready to be tested hard?
balance

Grade

G, test and score
Upgrade

What is analysed

  • arrow_right A comprehensive professional prior art search against your real claims
  • arrow_right The patentability assessment
  • arrow_right A mock examination that attacks your claims like a real office action
  • arrow_right The cumulative project scorecard

The crew on it

Cora, deep search Yuki, mock exam Noor, strategy Sean
door_frontGate to Execute. Prior art reviewed, patentability accepted, the examiner's objections resolved. Decision: is it strong enough to file?
rocket_launch

Execute

E, deliver the pack
Upgrade

What is produced

  • arrow_right Claims finalised, and the prior-art disclosure compiled where the office requires it
  • arrow_right A jurisdiction-formatted filing pack
  • arrow_right The plain-English reading and the attorney binder

The crew on it

Remy, final draft Iris, binder Noor, route and cost
checkDone. The filing pack is assembled, ready for your attorney or self-filing.

Not every project ends in a patent. For an improvement you would rather keep secret, IPGuru can take you to a trade-secret completion instead, documenting and protecting it without a public filing. Other objectives, such as open source, are on the roadmap.

The gates

What a gate really is

A gate is not a wall. It is a short, honest checkpoint between two phases: a checklist, a readiness read, and a decision that is always yours.

checklist

A checklist, not a queue

Each gate is a few required artefacts, shown plainly as "2 of 3 done". Never a ticket system to manage.

speed

A readiness read

The project scorecard gives an honest signal, in the journey header, of how ready you are to advance.

alt_route

A decision: go, loop, or pivot

You advance, iterate back for more work, or change direction. You are never forced forward.

edit

An override, with a reason

You can override a soft gate with an explanation. It is logged and carried into the application.

forum

Nova narrates it

When a gate is met, Nova summarises what was achieved and proposes the next step. You confirm with one tap.

lock_open

One gate is the upgrade

The gate out of Observe is where a free, validated idea becomes a paid project and the build unlocks.

The engine of quality

The Refine and Grade loop

Refine and Grade are not a one-way street. A weak grade sends specific direction back into Refine: a thin embodiment count, a claim exposed by prior art, an objection from the mock examiner. Each pass hardens the invention. You watch the claims get stronger, cycle by numbered cycle, until the examiner has no substantive objection left.

sync

How the loop turns

  • looks_one Grade tests the draft and names the two or three weakest points.
  • looks_two Refine strengthens exactly those, with the right agent on each.
  • looks_3 Grade re-tests. Repeat until it holds, then move to Execute.

Gold, autonomous iterative refinement

Set a goal, let the Virtual Inventor iterate

On Gold projects you can hand a gate to the Virtual Inventor, an AI that acts on your behalf. You set the goal and what success means; it does the iterative work to get there. You review the progress, and you make the gate decision.

flag

Goals that mean something

Set your own success criteria, or pick a standardised set that matches your strategy: optimisation, new feature, blending products, or category inspiration. A goal is tied to a gate, for example:

  • check_circle Reach product feasibility and viability.
  • check_circle Meet the novelty and inventive-step bar.
  • check_circle Clear the mock examination with no substantive objection.
smart_toy

The Virtual Inventor, on your behalf

It runs the refine loop autonomously toward your goal:

  • looks_one You set the goal, define success, and set any limits.
  • looks_two It generates, scores, researches, strengthens, and re-tests, cycle after cycle.
  • looks_3 It reports progress and pauses when it needs you or hits a limit.
  • looks_4 You review and make the gate call. You reach the gate; it does the work to get there.

You stay in control of the two things that matter: you define what success means, and you decide when a gate is reached. The Virtual Inventor never advances a gate on its own, and a PIVOT is always yours to make. Gold projects only.

u_turn_right

The five moves of a PIVOT

  • pause_circle Pause, and suspend the project's other work.
  • search Inspect what you have learned so far.
  • fact_check Verify the new direction makes sense.
  • explore Orient toward the stronger use, market, or claim.
  • move_up Transition: archive this project and seed a new one.

Changing direction

PIVOT, without losing work

Sometimes the evidence says the angle is wrong. Vega, the crew's coach, coordinates a PIVOT the way Nova coordinates the project. It is a whole-project event: your current project is archived as a read-only record, a new one is seeded from it, and your grade carries forward at no charge. Nothing is lost, and the history of why you changed course is kept.

Who does what

The crew across the journey

PhaseThe analysisWho leads it
FocusIdea capture, first concept imageNova, Kira
ObserveMarket, competitors, early patentability, prior artCora, Vega, Sean
RefineImaging loop, description, draft, claims, embodimentsKira, Remy, Sean, Iris
GradeDeep prior art, patentability, mock examination, scorecardCora, Yuki, Noor, Sean
ExecuteFinal claims, disclosure, filing pack, plain-English, binderRemy, Iris, Noor

Nova coordinates throughout; Sean is the lead author of your deliverables; the patent team (Noor, Remy, Yuki) joins on an upgraded project. Meet the crew.

Good decisions, in the right order.

Start with the free phases and let the evidence tell you whether to build.

auto_awesome Start free
No credit card to validate.