The verdict before the ledger.
JIBO evaluates a domain's total brand identity — how it looks, reads, and sounds — then anchors it in verified sales data. No outdated formulas or fake comps.
Anyone can build something that always returns a number. The harder engineering problem is a system that knows when it should stay quiet. These rules load before any valuation runs, and they can't be overridden by a persuasive story.
A comparable sale enters the pool only with a price, a date, and a venue attached. Single‑source reports are admitted at reduced weight and labelled as such. Figures that contradict the verified record are purged, not averaged.
Automated registrar valuations carry no comparable evidence and no method you can audit. JIBO treats those platforms as places to sell a domain — never as a source of what it's worth.
Marketplace listings tell you what a seller hopes for. They enter only as directional sentiment, discounted toward a realistic close, and they are never permitted to anchor a valuation.
Below a minimum count of class‑correct comparables, confidence is capped at Low regardless of how good the domain looks. The band moves with the evidence, not with the ambition.
A universal one‑word dictionary sale cannot anchor a coined brandable or a two‑word descriptive. Each class carries its own comparable pool, and the boundary is enforced in code.
Some extensions have no documented secondary market. For those, JIBO returns Speculative and names the reason rather than manufacturing a figure from a single historical outlier.
A common failure in domain appraising is comparing a name to the wrong sales comps. To prevent this, JIBO classifies the structural and linguistic properties before determining the domain price. This classification system is extension‑agnostic. The ranges below serve as the .ai reference basis, while other extensions (.com, .io, .xyz, .org, etc.) are priced independently against verified comps within their own TLD.
| Class | Definition | .ai reference range | Liquidity |
|---|---|---|---|
| A1 | Universal consumer or tech word — recognised in every language | $400K – $1.2M | Highest |
| A2 | Strong category word — clear read, narrower than A1 | $150K – $500K | High |
| A3 | Dictionary word — niche, emotional, or abstract | $75K – $200K | High |
| B1 | Phonetic equivalent of a dictionary word | Fraction of the word | Medium |
| B2 | Strong coined brandable — pronounceable, tech‑adjacent | $25K – $80K | Medium |
| B3 | Abstract coined — pronounceable, no category read | $8K – $35K | Low–Med |
| C | Two‑word compound or abstract coinage | $5K – $25K | Low–Med |
| D | Descriptive phrase — functional words | $2K – $15K | Low |
| D · exc | Two‑word that reads as a natural brand in a hot vertical | up to $100K | Medium |
| E1 | Meaning‑anchored initialism | $100K – $250K | Medium |
| E2 | Brandable letters, multiple expansions | $15K – $60K | Med–Low |
| E3 | Random letters, no meaning | $2K – $10K | Low |
These don't run in parallel and they can't be reordered. A structure penalty applied after an uplift produces a different — wrong — answer than the same penalty applied before it. The sequence is the part most tools skip.
Asset class and sub‑tier assigned. Everything downstream inherits this decision, so it happens first and nothing proceeds until it resolves.
Eight weighted factors — brandability, clarity, pronunciation, length, extension fit, legal exposure, vertical fit, onchain fit — position the name within its class rather than against the whole market.
Class‑correct comps retrieved in strict priority order, then weighted for venue quality and recency. Wholesale and expired‑auction results are context only, never retail anchors.
Word count, compound quality, extension redundancy, and phonetic‑variant drag applied here — before any uplift can compound them.
A hard cap. If no realistic buyer persona for this name carries the budget, the valuation is reduced to what someone will actually pay, whatever the comps suggest.
Gated on validation, not narrative. No live protocol earns the minimum multiple; confirmed revenue or a signed commitment earns more. Unvalidated stories earn nothing.
Payments and AI infrastructure clear faster than geographic or abstract niches. The adjustment reflects time‑to‑sale, not desirability.
Buy‑it‑now and make‑offer floor derived from the tier matrix, with psychological pricing that matches the band — precise below fifty thousand, institutional above it.
The listing format named — buy‑it‑now versus make‑offer, because running both on the same asset undercuts each. Where you take it — the marketplace of your choosing, a live DARWA deployment, or both at once — stays your call.
A valuation without a confidence band is a guess wearing a suit. JIBO publishes the band with the number, every time, and the band is set by the evidence available — not by how much you'd like it to be.
High
±10%
Deep, recent, class‑correct comparable support. The number is defensible in a negotiation.
Medium
±20%
Adequate comps. Also the hard ceiling whenever a meaningful share of the value rests on narrative.
Low
±40%
Thin market data. Treat as a starting position for outreach, not a listing price.
Speculative
±60%+
Insufficient evidence to appraise. JIBO names the gap instead of filling it.
JIBO prices the asset. The rest of the GhostFire stack deploys it, gives it a voice, and takes it on‑chain. Each engine runs standalone — together they cover a domain's entire life from acquisition to exit.
JIBO
The Oracle put to work. Researched, analyzed, and comp‑anchored. Market‑studied, classified, audited, and upgrade‑instilled.
You are hereDARWA™
The domain is the input. The identity is the output. Transforms a domain into cinematic identity, in seconds. No prompt. No templates. Just the name.
darwa.ai → IdentificationAGENTTWIN™
The Identity Layer For Autonomous AI. Every Agent needs an identity. Every identity needs a twin. A brand‑voiced concierge inside the deployment.
agenttwin.ai → TokenizationTOKENUP™
A GhostFire internal system for turning domains into Domains As Real-World Assets. A structured process for evaluating, preparing, and routing digital real estate.
tokenup.ai →No account, no prompt, no questionnaire. The engine classifies the name, pulls its class‑correct comparables, and returns a buy‑it‑now, a make‑offer floor, the likely buyer, and an honest confidence band.