Chat that thinks while it talks.
While one model writes your answer, a second one reads it as it comes back. That second model is Jev, a decision model: it never writes a word. It reads each sentence as it arrives and classifies it: the substance you asked for, extra detail you didn't, or padding such as filler, hedging and repetition. You can see at a glance what is worth reading. Before a word is written it also decides which model should answer and how long the answer should be, and afterwards which follow-up questions to offer. Each decision takes a fraction of a second, and Reflex shows you every one.
What Jev decides on every turn
In one call, Jev reads your message and makes two decisions. Which model should write: the cheap one, or, if it is unsure, the big one (about ten times the price)? And how long should the answer be: brief (a sentence or two), short (a paragraph) or detailed (steps, code, trade-offs)? The writer starts only once both are settled, with the length in its instructions and a matching limit. When Jev is torn between two lengths it picks the longer, so answers are never cut short to save money. Code always gets the full limit.
Each sentence is labelled as it arrives, three per call: substance (what you asked for), extra (true and related, but more than you asked), or padding (filler, hedging, repetition). Once the answer is complete, Jev also picks the one key sentence and it is highlighted. The answer restyles itself as you read, and a "Just the answer" switch hides the extras and padding altogether, telling you how much shorter that makes it.
With every batch Jev is also asked whether your question is now fully answered. At 90% sure, the stream is cut. In practice this rarely fires, because the length is already set in step 01: see the note below.
The writer drafts five follow-up questions; Jev scores them and keeps the best three. Selecting is a decision, so Jev does it. Writing is not, so it doesn't.
Jev does not measure carbon: the estimate is arithmetic on token counts. But its decisions shape the footprint and explain it. Routing keeps most turns on the small model, and the difference from always using the big one is shown as carbon not emitted. Setting the length up front means fewer tokens are written at all. Its sentence labels show how much of each turn's footprint went on padding rather than on the answer. And Jev's own calls are counted in the total, so its overhead is never hidden.
Is it fast enough to cut an answer short?
Partly. A verdict takes 0.3 to 0.6 seconds to come back, and the writer streams roughly a hundred tokens a second, so thirty to sixty more tokens appear before a cut can land. Reflex keeps any whole sentences from that window and drops a half-written one, so a stopped answer never ends mid-sentence.
The bigger limit is not speed. Across the test questions we ran, Jev only became 90% sure an answer was complete once the writer had, in fact, finished. The writer is told to be concise, so there is rarely a tail of recap to cut. Early stop is a safety net for answers that ramble, not the main saving. When it does nothing, the app says so rather than claiming a stop.
Cutting an answer short after the fact rarely works, so Jev now sizes it before it starts. Waiting for that decision adds about a third of a second, which is small next to the one to four seconds the writer takes to produce its first word. On our ten two-part starter questions, answers came out about 38% shorter in tokens than before (1,654 down to 1,024, one run of each, so treat it as indicative). A simple fact such as the chemical symbol for gold now comes back in a handful of tokens.
A decision model from TypeSafe. You give it some text and a typed question (pick one of these, yes or no, rate on this scale) and it returns the answer with a probability. It cannot write a sentence.
About a quarter of a second per call once a connection is open, up to 0.7 s on a cold one. Its servers are in Oregon; most of that time is the round trip from Europe, not the model.
$0.042 per million input tokens, output free. A typical turn spends less on Jev than a hundredth of a cent, but that is often more than the cheap writer's answer cost. Judging every sentence is not free, and the app shows it.
Everything is on the table
Beside the chat, the Flow view draws how Jev fits around the writer and lights up each step as your turn passes through it: the model and length it chose, how it labelled the sentences, whether it stopped the answer, the key sentence and the follow-ups. Below that, every turn has a card with each decision, its latency and the probabilities behind it. Every answer carries its token count, its cost and an estimate of its carbon. The conversation is also drawn: a tree that grows a branch per turn and a leaf per useful sentence, above layers of earth whose thickness is the estimated CO₂ each turn cost. The carbon figures are rough ranges built from published estimates, and the app explains how.
Writer: z-ai/glm-5.3-flash, escalating to z-ai/glm-5.3, via OpenRouter. Decisions: Jev (jev-latest) on TypeSafe's API. Costs are as billed; comparisons price the same tokens on GLM 5.3 with no decision layer.