Roadmap

Ordered by what unlocks the next thing, not by what demos best. No dates — dates on a one-person project are decoration.

---

Shipped

The landing page renders live from both upstreams with honest degradation. Six API endpoints serve the same assembler the page uses. Session engine with holiday handling, the ERC-8056 multiplier applied in exactly one place, the readout with sorting and per-row pool ladders, the execution check with depth guards, the verify drawer, the canvas pool field, and the design system as CSS custom properties with /system rendering the real components.

Next

---

What might not ship

Documentation that only describes the happy path is marketing. These are genuinely uncertain, with the actual reason rather than a hedge.

Continuous executable quotes — may not be possible

A real net figure needs a quoter call per pool, per size, per token. Roughly 194 tokens × ~20 pools × 4 sizes is 15,000 calls a minute, which no public RPC will serve and a paid one prices well above this project's budget.

Likely outcome: quotes become on-demand only, computed when someone asks for a specific token and size, with the continuous feed staying gross-and-labelled. /v1/quote may never have a batch equivalent.

Uniswap v4 pool keys — hard, not blocked

Quoting a v4 pool requires its PoolKey: currency pair, fee, tick spacing, hooks. Screener APIs return an opaque pool id instead, so the keys have to be recovered by indexing Initialize events from the PoolManager and maintaining that index forever. Solvable, but it is the largest piece of engineering in the plan and everything above sits behind it.

The signal itself could shrink — market risk

The weekend dislocation exists because nothing makes a market when the reference is frozen. That is already being addressed: a launchpad said publicly it approached market makers about weekend coverage, and the issuer is reported to be adding liquidity. If they succeed, the most dramatic part of this product gets smaller. The pool-to-pool and cross-pair work survives; the 35× headline probably does not repeat.

The reference API has no contract — dependency risk

Keyless, undocumented as to stability, no service commitment. It could be gated, throttled or moved without notice. Fallbacks exist — the on-chain Chainlink feed covers tokens that have one, a commercial equities API covers the rest — but a free product with a paid dependency is a different product.

Real-time websocket — later

Persistent connections at any scale need more than one small server. It ships when there are enough holder-tier users to justify it. The 60-second revalidated endpoint is the honest default.

The on-chain oracle — may never ship

Engine 03 in docs/token.md is the most valuable idea here and the least certain. It needs contracts, a staking and dispute mechanism, and — the real gate — a protocol that actually wants to read it. If no launchpad or lending market integrates, building it would be infrastructure cosplay. It ships when a named consumer asks, and if none does, this section is where that gets recorded.

Cross-chain coverage — later

Tokenised equities exist elsewhere; Solana's xStocks carry real liquidity and the same basis logic applies. Different chain, different pool infrastructure, different reference API. Out of scope until the first chain is genuinely finished.

---

Standing commitment

Corrected 2026-09-03 — two roadmap items were on the landing page in the present tense

Found while restructuring the page, under the heading How you would catch us lying, which is the worst place on the site to overstate anything.

• Hash-chained samples (v0.4). The page read "Each sample stores the hash of the one before, and the head is anchored on chain daily. Every minute missed is logged with its cause, in public." Nothing persists between requests in this build. Now stated as a plan, flagged NOT BUILT, with the gate named.
• The weekend board (v0.6). The page read "Every Friday at 20:00 ET the frozen reference is published. Every Monday, what the pools did against it." No board exists and nothing has been posted. Same treatment. The footer separately advertised "Weekend report · Mondays 09:00 UTC", which nothing schedules; that line is gone.
• The on-chain oracle (Engine 03). The token section showed only GATE 90 days of history, which reads as a schedule for something on its way. This section says it may never ship, and the real gate is a protocol that wants to read it. Both now appear on the card.

Standing check, since the phrasing is easy to reintroduce:

grep -rn "Each sample stores\|the frozen reference is published\|Weekend report" src/   # must be empty

Corrected 2026-09-03 — the landing page implied imminence it had not checked

The corporate-action panel filtered on status === 'IN_PROGRESS' and captioned the result "Each one pauses that token's oracle while it processes" — the exact mistake docs/data-sources.md warns about. On the day it was found it listed dividends 13 to 28 days out. It now renders the desk's ActionCalendar, which separates the 14-day window from the 2-day imminence badge. One component, one behaviour, both pages.

Anything in this section that turns out to be impossible gets marked as such here, with the reason, rather than quietly disappearing from the roadmap. A roadmap that only ever grows is a roadmap nobody should trust.