Plain-language glossary
Jev AI model glossary
Every term you meet while building with Jev, defined in plain language and grounded in TypeSafe's documentation: the three primitives, confidence and calibration, model IDs and pricing, the named patterns, and the jagged edges the docs admit to.
Each entry links to the resources and docs behind it. Definitions track the current jev-1.13.0 documentation, last reviewed .
- Terms
- 177
- Categories
- 7
- Linked resources
- 142
- Docs links
- 136
Browse by category
7 groups, from core concepts to the ecosystem around the API.
- Core Concepts21The vocabulary behind System One: what Jev is, what state and questions are, and the architecture TypeSafe recommends.
- Primitives17Choice, Score, and Noul, plus the instructions, criteria, and answer fields you use to define them.
- Answers & Confidence14How probabilities, confidence, calibration, and thresholds turn typed answers into decisions code can gate.
- API & Models31The endpoint, model IDs and aliases, context limits, token pricing, retries, and error handling.
- Patterns35The named compositions the docs and cookbooks use: fan-out, confidence gates, cascades, re-ranking, and more.
- Evaluation & Training40RLCD and its alternatives, the documented jagged edges, and how Jev's quality is measured.
- Ecosystem19The SDKs, skills, MCP servers, gateways, and community ports that put Jev inside your stack.
All 177 Jev AI model terms
Search the glossary, filter by category, or jump to a letter.
177 terms
A
- Adversarial contentPrompt injectionJev treats state as data, not as hostile input; injected instructions can move answers, so write precise criteria and test edge cases.Evaluation & Training
- Agent loopLLM agentAgentic loopA control pattern where the model chooses its next step repeatedly; TypeSafe's guidance is to prefer a code-owned workflow instead.Core Concepts
- Agent skillTypeSafe agent skillA drop-in SKILL.md that teaches coding agents to structure Jev states and questions and to batch questions per call.Ecosystem
- AI gatewayModel gatewayModel routerA routing layer that serves Jev alongside other models; Vercel AI Gateway lists typesafe-ai/jev as a hosted model.Ecosystem
- AI primitivePrimitiveOne of TypeSafe's small, typed building blocks (Choice, Score, Noul) that pair a question with a typed answer you compose in code.Core Concepts
- AI-powered softwareThe architecture TypeSafe recommends: code owns the workflow and the model appears only where programmable common sense is needed.Core Concepts
- AnswerTyped answerThe structured result returned under a question's ID: a choice, a score, or a probability of yes, plus distributions and confidence.Core Concepts
- API error529 OverloadedThe four documented failure statuses (401, 422, 429, and 529), with retry guidance for the last two.API & Models
- API keyTYPESAFE_API_KEYThe bearer credential for the TypeSafe API, created in the console and read from TYPESAFE_API_KEY by the official SDKs.API & Models
- AssetsAssets mapImage assetsThe top-level request map that carries images into a Jev call: each image is keyed by a reference string and placed anywhere the JSON interface allows.API & Models
- Atomic questionA question that asks one specific, well-scoped thing, the kind of judgment a knowledgeable person makes in a few seconds.Core Concepts
- AutoresearchAutoresearch feature discoveryA loop that proposes questions, turns free text into numeric features, and uses model errors to improve a classical regressor.Patterns
B
- Backtick path referenceDot-and-index pathNaming a state field inside instructions with backticked dot-and-index paths, such as `ticket.messages[0].text`.Primitives
- Base ratePrior probabilityPrevalenceHow common the positive class is in your traffic, which sets the starting probability a decision model has to move away from.Answers & Confidence
- Base URLTYPESAFE_BASE_URLThe host a Jev client calls. Point it at a local System One server and the rest of the request stays the same.Ecosystem
- Benchmark claimsVendor benchmarksTypeSafe's headline multipliers (193.6x faster, 444.6x cheaper, a 238x lower input price) come from its own workflow comparisons.Evaluation & Training
- BoxBox questionBounding boxThe experimental Jev 1.14 question type that answers where something is in an image: four edges as fractions of the image, plus a confidence.Primitives
- Brier scoreA score for a full probability vector against one observed label. Lower is better, and a confident wrong answer costs more than a hedge.Evaluation & Training
C
- CalibrationProbability calibrationThe property that outcomes given a probability of 0.2 occur about 20% of the time, measured across many predictions.Answers & Confidence
- Canary rolloutCanary deploymentActing on the new model for a small slice of traffic and watching the outcome metric, with a one-flag rollback to the incumbent.Patterns
- Candidate coverageCoverageThe share of real inputs for which the option list actually contains the right answer, measured separately from the model's accuracy at picking it.Evaluation & Training
- ChoiceChoice questionPick one option from a set you define; returns the selected option, a probability for every option, and confidence.Primitives
- Choice optionOptionOne key in a Choice criteria map, with a description that can be a string, a structured object, or null.Primitives
- Circuit breakerBreaker patternA client-side state that stops sending requests after repeated failures, then probes tentatively before closing again.Patterns
- Client SDKTypeSafe SDKThe official typed clients for the API (Python and JavaScript/TypeScript), with retries and inferred answer types.API & Models
- Coding agentAI coding agentThe fastest-growing corner of the Jev ecosystem: MCP servers, tool-call gates, context pruners, and skill rankers.Ecosystem
- Community portCommunity SDKAn unofficial client or integration for a language or framework TypeSafe does not ship officially, from Go and .NET to DSPy.Ecosystem
- Composable AITypeSafe's manifesto position: intelligence as a dependable primitive that software branches on, rather than an assistant that keeps humans in the loop.Core Concepts
- Composite scoringWeighted scoringBreak a judgment into atomic Score questions, normalize each, and combine them with weights you control in code.Patterns
- ConfidenceConfidence scoreA 0-1 number on every Choice and Score answer that collapses the probability distribution into one certainty score.Answers & Confidence
- Confidence floorMinimum confidenceA baseline below which code never acts on an answer, regardless of what the answer says.Answers & Confidence
- Confidence gateConfidence thresholdA threshold on confidence that decides whether code acts, confirms first, or escalates to a person or a bigger model.Answers & Confidence
- Confidence-gated routingConfidence routingUse the answer to decide what to do and confidence to decide whether to act, with per-action thresholds set by the cost of being wrong.Patterns
- Confusion matrixError matrixThe table of predicted versus actual labels that shows which classes a model actually confuses, not just how often it is right.Evaluation & Training
- Consensus labelsGround truthReference answers built from two frontier models at high thinking, used as the ground truth in TypeSafe's workflow evals.Evaluation & Training
- Context rotThe accuracy loss that comes from padding state with detail unrelated to the question; filter state before sending it.Core Concepts
- Context window64k contextThe 64k-token budget covering state plus all questions, with a 32k sub-budget for state plus the longest question.API & Models
- Contradictory criteriaWhen instructions and criteria ask for different things, accuracy drops; treat criteria as an extension of the instruction.Evaluation & Training
- CookbookAn end-to-end recipe in the TypeSafe docs that shows a real problem, with its dataset, latencies, and measured numbers.Ecosystem
- Cost matrixUtility matrixLoss matrixThe table of what each kind of error costs, which turns a probability into the threshold that minimizes expected cost on a given action.Answers & Confidence
- Cost/quality frontierPareto frontierPlotting accuracy against cost and time on log scales, where up and to the left is better; cascades are designed to sit there.Evaluation & Training
- CriteriaRubricThe possible answers for a question: a map of options for Choice, an ordered list of levels for Score, or true/false for Noul.Primitives
- Cross-entropy lossLog lossCross-entropyThe training objective that penalizes a confident wrong probability monotonically, and the proper scoring rule behind log loss in evaluation.Evaluation & Training
D
- Danger zoneDanger-zone headerDreamer-readyTypeSafe's name for opt-in API headers that enable experimental Jev features (images, Box questions, long context) that it will not yet vouch for in production.Patterns
- Dates as textDate comparisonJev reads dates as strings, not ordered quantities; extract components with Choice and do the ordering and math in code.Evaluation & Training
- Decision headThe small trainable module added to a frozen backbone that turns hidden states into one probability per option.API & Models
- Decision hijackingProgram decision hijackingContext attacksAn attack that steers a typed decision through attacker-controlled context (tool output, retrieved text, or an option name) while every request and response stays schema-valid.Patterns
- Decision IndexJev Decision IndexHF Decision IndexA Hugging Face leaderboard that scores open reproductions of Jev on the same frozen decision suite.Evaluation & Training
- Decision modelSystem One decision modelA model that returns a typed decision (a choice, a score, or a probability) instead of generated text.Ecosystem
- DecompositionBreaking a broad judgment into one question per factor, then combining the answers with weights and rules you control in code.Core Concepts
- DistillationTeacher-student distillationTraining a smaller model to reproduce a larger model's output distribution, including its probabilities rather than labels alone.Evaluation & Training
- Distribution shiftData driftCovariate shiftWhen the traffic a model sees at run time stops looking like the data it was calibrated on, so its probabilities no longer mean what they did.Evaluation & Training
- Document splittingAsking Jev where one packet becomes the next document, then letting code export the page ranges.Patterns
- Drop-in model swapTYPESAFE_MODELPointing an existing Jev client at another model or host without rewriting the question or answer types.API & Models
E
- Encoder backboneThe bidirectional transformer under a decision model that reads the state and options once, rather than generating tokens.API & Models
- Entity alignmentEntity matchingsameAsDeciding whether two records describe the same entity, with one Score per pair plus companion Nouls for the fields that disagree.Patterns
- Entry typeField typeThe value shapes accepted throughout a request (string, object, array, or null) for state, instructions, and criteria fields.Primitives
- Expected calibration errorECEThe average gap between predicted confidence and observed accuracy, grouped into bins. Lower is better, and the number depends on the bins.Evaluation & Training
- Exponential backoffThe retry strategy the docs recommend for 429 and 529 responses: wait longer after each attempt instead of retrying immediately.API & Models
F
- Fail closedDeny by defaultWhen the model errors, times out, or lacks evidence, the system escalates or blocks instead of allowing the action.Ecosystem
- Fail openAllow by defaultWhen Jev is down or slow, the gate allows the action, including rules that would otherwise block it.Ecosystem
- Fine-tuningContinuing training on your own labelled decisions, which is where an open checkpoint stops being a proof of concept.Evaluation & Training
- Function callingTool callingTurn natural language into typed function calls: closed-set arguments become Choice questions and optional ones become Nouls.Patterns
G
- Generation limitationText generationJev is not trained to generate text; forcing it by chaining choices is slow and unreliable, so extraction becomes a Choice instead.Evaluation & Training
- GuardrailContent moderationScreening messages going into or out of an LLM app with hazard Nouls and severity Scores, then thresholding the probabilities.Patterns
H
- Hierarchical classificationDeep taxonomyWalk a deep taxonomy level by level, using each Choice answer to pick the next level's options, keeping several paths alive.Patterns
- Human reviewEscalationThe explicit path for uncertain or high-stakes cases: route to a person instead of acting on a low-confidence answer.Patterns
I
- IdempotencyIdempotency keyIdempotent requestThe property that repeating a request produces the same effect, which is what makes a retry safe to send after a timeout.API & Models
- Inbox triageEmail classificationOne Jev call per message, with a threshold and an action per label, so code moves or flags mail the model only scores.Patterns
- IndirectionMulti-hop questionsQuestions about a property of a property, or with double negatives, cost accuracy; write instructions as directly as possible.Evaluation & Training
- Input tokensinput_tokensThe tokens billed on every request: state plus questions, including speculative questions your code ignores.API & Models
- InstructionsThe field that states what a question should judge; accepts a string, an object, or an array for extra structure.Primitives
- Intent routingIntent classificationClassify intent with a Choice and complexity with a Score, then send each branch to deterministic code, a specialist model, or a person.Patterns
- Isotonic regressionIsotonic calibrationA non-decreasing fit from raw confidence to observed accuracy. It can follow a miscalibration that is not a sigmoid.Evaluation & Training
J
- JaggednessModel jaggednessThe documented set of jev-1.13 failure modes (literal reading, counting, dates, indirection, context rot), each with a guardrail.Evaluation & Training
- JevJev AI modelJev (AI model)jev-latest modelTypeSafe AI's flagship AI model and the first System One model: typed decisions and calibrated probabilities instead of generated text.Core Concepts
- Jev EngineeringThe practice of splitting a workflow into an LLM that writes, Jev that decides, and code that acts.Ecosystem
- jev-1.13.0Jev 1.13The current versioned Jev build: $42 per billion input tokens, a 64k context, and 250,000 tokens per second.API & Models
- jev-1.14.0Jev 1.14Jev Day releaseThe October 14, 2026 Jev build behind jev-preview: small improvements everywhere, the largest expected in multilingual support, plus the only model that answers Box questions.API & Models
- jev-latestJev latestThe default model alias in the SDKs and docs: the most recent stable official Jev release, currently jev-1.13.0.API & Models
- jev-previewJev previewThe alias for the most recent Jev release whether or not it is official; currently identical to jev-latest.API & Models
- JevBenchJev BenchAn independent benchmark for Jev-class decision models, scored as the harmonic mean of intelligence, calibration, speed, and cost.Evaluation & Training
K
L
- Language supportMultilingual supportEnglish is Jev's primary training language; other languages are accepted but less accurate, so test before relying on them.API & Models
- Latency percentilesp95 latencyTail latencyReporting round-trip time as p50, p95, and p99 rather than an average, because the tail is what a control loop feels.API & Models
- LeaderboardA public ranking of decision models on a shared task set, where the selection rules decide what the numbers mean.Ecosystem
- Line-by-line searchSemantic findRank every line of a document in one request with a Choice, while a Noul checks whether the document contains an answer at all.Patterns
- Literal readingJev answers the question you wrote, not the one you meant; scoping words, negations, and implied conditions are read at face value.Evaluation & Training
- Local decision serverLocal model serverA process on your machine that answers typed questions from local weights, so the state does not leave the host.Ecosystem
- Logit readoutLogit scoringReading the next-token probability of each declared label, then renormalising, so no answer text is generated.Patterns
- Long context262k contextDanger-zone long contextThe opt-in mode that raises both Jev context caps to 262k tokens behind the TypeSafe-Danger-Zone-Long-Context header, at a reliability cost TypeSafe warns about.API & Models
M
- Machine Native IntelligenceTypeSafe's name for AI with software-like properties: structure, reliability, observability, testability, speed, consistency, and low cost.Core Concepts
- Machine-to-machine automationM2MAI-to-AIAutomation where software calls AI directly, expected to be roughly 99% of large-scale AI interactions versus 1% human-facing.Core Concepts
- MCP serverModel Context Protocol serverA server that exposes Jev to coding agents as a tool through the Model Context Protocol, usually one evaluate call.Ecosystem
- Memory gateRemember gateA second Noul question that asks whether a request states a fact worth remembering, so a wrapper can save it into the state of that user's later calls.Patterns
- Mode droppingMode collapsePreference optimization narrowing a model's output distribution toward a favored style and away from other valid outputs.Evaluation & Training
- Model aliasA name like jev-latest that resolves to a versioned model ID and can move when a new release ships.API & Models
- model fieldThe required request field that selects which model handles the call; accepts versioned IDs and aliases.API & Models
N
- Neuro-symbolic AINeuro-symbolicThe manifesto's framing of TypeSafe's aim: neural networks for perception paired with symbolic logic, or smart if-statements.Core Concepts
- Non-autoregressive decodeOne forward passParallel scoringScoring every possible answer in a single forward pass instead of generating tokens one at a time.Core Concepts
- NormalizationScore normalizationScaling each Score by its number of levels (dividing by len(criteria) - 1) before weighting it in a composite.Patterns
- NoulNoul questionYes/no questionThe yes/no primitive: one probability from 0 to 1, where the value is the answer and no separate confidence is returned.Primitives
- Noul criteriaYes/no criteriaOptional true and false descriptions that define what a yes and a no mean for a Noul question.Primitives
- Noul probabilityP(yes)The single 0-1 value a Noul returns: the probability that the answer is yes, with no separate confidence field.Answers & Confidence
- Noul-only modelA local decision model that answers yes/no questions and rejects Choice and Score instead of guessing a shape.API & Models
- Null criteriaNull option descriptionPassing null as a Choice option description when the option name needs no extra explanation.Primitives
O
- ONNXA portable model format that runs outside Python, including in browsers through WebGPU and in on-device runtimes.Ecosystem
- Open weightsOpen-weight modelModel checkpoints published for anyone to download, run, and fine-tune, as opposed to a hosted closed API like Jev.Ecosystem
- Option markerScoring each answer option at its own marker token, then softmaxing over the question's options so the answer space is defined at request time.API & Models
- Option order biasOrder sensitivityOrder-flipThe change in a model's answers when the same options are presented in a different order, measured as an order-flip rate on reversed or rotated candidate sets.Evaluation & Training
- Option-label biasLabel-over-definition biasThe tendency of a typed decision model to answer from an option's label rather than the definition attached to it, so renaming options to A and B can raise accuracy.Evaluation & Training
- Oracle floorError floorThe best error rate achievable with perfect choices from the available options, used to judge how close a system gets.Evaluation & Training
- Output tokensoutput_tokensThe tokens Jev emits for answers; they are reported in usage and are free of charge.API & Models
P
- Parallel beam searchBeam searchKeep the top-k branches at each taxonomy level instead of only the winner, so a wrong first choice can still be recovered.Patterns
- Parallel questionsQuestion batchingBatching many questions into one request instead of many calls, which is dramatically cheaper and faster with identical answers.Patterns
- Per-tick decisionGame-playing agentOne structured game state and the legal actions, sent to Jev on a timer, with code applying or rejecting the pick.Patterns
- Platt scalingSigmoid calibrationA two-parameter sigmoid fit on labeled data that maps a model's raw probability onto a better-calibrated one.Evaluation & Training
- PlaygroundTypeSafe playgroundThe browser workbench for pasting state and typed questions and inspecting answers, probabilities, confidence, latency, and usage.API & Models
- Pre-parsed value extractionSpan extractionRegex finds candidate spans, then Jev picks the requested one so code copies a verbatim value that cannot be invented.Patterns
- Precision and recallPrecision/recallF1 scoreThe two error-facing rates of a classifier: how often a yes is right, and how often a real yes is caught, plus the F1 that blends them.Evaluation & Training
- Primitive limitsThe hard caps to design around: 255 Choice options, 2 to 10 Score levels, and a 64k context covering state plus questions.Primitives
- Probability coherenceCoherenceWhether a model's probabilities for logically linked questions fit together item by item: a label and its negation should sum to one, and single-label probabilities should not exceed it.Evaluation & Training
- Probability distributionprobabilitiesThe floats summing to 1 across your options or levels that every Choice and Score answer returns under probabilities.Answers & Confidence
- Proper scoring ruleStrictly proper scoring ruleA reward or error measure that is optimized only by reporting true probabilities: the training signal behind RLCD.Evaluation & Training
Q
- QuantizationStoring model weights at lower precision, 8-bit or 4-bit, to cut size and memory, at some risk to fine-grained probabilities.Ecosystem
- QuestionTyped questionA named judgment for Jev to make about the state, defined by an ID, a type, instructions, and criteria for the typed primitives.Core Concepts
- Question IDQuestion keyThe key you choose for a question; answers come back under the same key, and the ID is never sent to the model.Primitives
R
- RAG passage classificationPassage filteringScore each retrieved passage with four Noul questions, then decide in code which passages reach the answering model.Patterns
- Rate limit429 Too Many RequestsThe throughput caps (250,000 tokens per second and 1,200 requests per minute) that return 429 when exceeded.API & Models
- Re-rankingRerankingScore each query-candidate pair with one question and reorder a shortlist; the CLERC example lifts top-1 accuracy from 5% to 18%.Patterns
- Rejection optionNone-of-the-above optionOther optionAn explicit choice like other or none-of-the-above that gives the model a way to decline, and the reason an option list without one can only pick the least-wrong answer.Primitives
- Reliability diagramCalibration curveThe plot that bins predictions by confidence and compares each bin's average probability to the observed rate of correct answers.Answers & Confidence
- RepeatabilityProbability standard deviationHow stable answers are across repeated evaluations, measured separately from accuracy and reported as probability standard deviation.Evaluation & Training
- Retry policyRetryPolicyThe SDK configuration for retrying failed requests: attempt count, retryable statuses, backoff, and retry-header handling.API & Models
- Risk triageJev scores a transfer or a claim. Hard rules and a routing matrix in code decide whether to pass, review, or freeze.Patterns
- RLCDReinforcement learning for calibrated decisionsTypeSafe's post-training method: reinforcement learning that optimizes for calibrated decisions and probabilities instead of generated text.Evaluation & Training
- RLHFReinforcement learning from human feedbackThe human-preference training method that turned pretrained models into chatbots, and the method TypeSafe deliberately did not use.Evaluation & Training
- RLVRReinforcement learning with verifiable rewardsTraining against automatically verifiable rewards, which produced strong reasoning models that are slower and more expensive.Evaluation & Training
S
- ScoreScore questionRate the state against 2 to 10 ordered levels; returns a probability-weighted position, per-level probabilities, and confidence.Primitives
- Score legendLegendThe Score answer field that maps each level number back to its description, so results stay readable after the fact.Primitives
- Score levelLevelOne entry in a Score's ordered criteria list; levels are numbered by array position from 0 and judged independently.Primitives
- SDE cascadeStructured-data-extraction cascadeA two-stage extraction cascade: a mini model extracts, a Noul battery verifies, and a confidence gate escalates to a reasoning model.Patterns
- Second requestA follow-up call that is only justified when code cannot build the second request until it has the first answer.Patterns
- Selective decisionReject optionSelective predictionA decision that is allowed to abstain instead of answering, trading coverage for accuracy on the cases it does handle.Answers & Confidence
- Selective escalationAccept-or-escalateSending only the decisions a fast model is least sure about to a slower, stronger system, so most traffic is answered cheaply and the hard tail still gets quality.Patterns
- Self-consistencyRepeatability across repeated evaluations: Jev's probability standard deviation measured at about 0.01 in the consistency cookbooks.Patterns
- Shadow modeShadow deploymentRunning a new model beside the incumbent and logging its answers without acting on them, to learn its distribution before trusting it.Patterns
- Skill suggestionSkill rankingRank a large skill catalog in one request, then re-read the top few in full detail and pick at most one.Patterns
- Smart if statementSmart if-statementTypeSafe's shorthand for what a System One decision is: a condition your code can branch on directly, backed by a model judgment.Core Concepts
- Speculative fan-outFan-out patternAsk every question the system might need in one request (including ones only some branches use), and let code discard the rest.Patterns
- Speculative questionA question whose answer only matters on some code paths, asked anyway because questions are cheap and parallel.Patterns
- StateThe content a request asks Jev to evaluate: a string, a named object, or an array of messages, shared by every question in the request.Core Concepts
- Structural invariantsComplementary probabilitiesArithmetic identities you might assume (complementary probabilities summing to 1, agreement across question types) are not guaranteed.Evaluation & Training
- Structure recoveryAutoformatReconstruct Markdown from unformatted text in two requests: stitch split lines, then classify every block's type.Patterns
- Structured instructionsStructured criteriaPutting the question in one field and its supporting data, examples, or focus notes in others instead of one long string.Primitives
- SycophancyPeople-pleasingThe tendency of preference-trained models to say what people want to hear, one of the reasons TypeSafe chose a calibration objective.Evaluation & Training
- System OneSystem One modelThe class of AI models built to make fast, structured decisions software can use directly, named after Kahneman's fast-thinking System 1.Core Concepts
- System One adaptersystem-one-adapter-pythonThe official drop-in TypeSafeClient replacement backed by OpenAI, Anthropic, and compatible LLM APIs.Ecosystem
- System One endpointPOST /v1/systemoneThe single HTTP endpoint that evaluates a state against typed questions: POST https://api.typesafe.ai/v1/systemone.API & Models
- System Two taskSystem 2 taskWork that needs extended, multi-step reasoning; the docs treat it as a signal to decompose rather than a job for Jev.Core Concepts
T
- Temperature scalingTemperature fittingDividing a model's logits by one fitted number per setting to bring its probabilities closer to observed frequencies.Evaluation & Training
- Text-only inputText-onlyJev evaluates strings, JSON objects, and arrays of text; images, audio, and video are not supported and must be pre-processed.Core Concepts
- The Bitterest LessonThe Bitter LessonTypeSafe's essay extending Sutton's lesson: doing the right task beats data, which beats compute, which beats algorithms.Evaluation & Training
- ThresholdDecision thresholdThe value your code compares against a score, a Noul probability, or a confidence to turn an answer into an action.Answers & Confidence
- Token pricingBtokJev's published rate: $42 per billion input tokens, with output tokens free.API & Models
- Typed decisionStructured outputA model output that conforms to a software type your code expects, rather than free-form text that has to be parsed.Core Concepts
- TypeSafe AITypeSafeThe company behind Jev and the System One API, including the official SDKs, cookbooks, console, and published workflow evals.Core Concepts
- TypeSafe Consoleconsole.typesafe.aiThe web console for API keys, model management, and the playground, with walkthrough lessons and realistic use cases.API & Models
U
- UncertaintyAbstentionThe honest "I'm not sure" signal that lets a system defer instead of guessing, exposed through probabilities and confidence.Answers & Confidence
- Uncertainty bandReview bandA middle range of probabilities, such as 0.30 to 0.70, that maps to an explicit uncertain outcome instead of a hard yes or no.Answers & Confidence
- Unofficial hostIndependent Jev hostJev resellerAn independent site that wraps Jev's API with its own keys, playground, or prepaid credits.Ecosystem
- Unreliable countingCountingJev does not count reliably (characters, occurrences, or list items), and error grows with the size of the thing counted.Evaluation & Training
- UsageToken usageThe response field reporting input and output token counts for a request, used for cost tracking.API & Models
W
- Weighted scoreScore expectationThe Score answer computed as each level number times its probability, which is why a score can land between two levels.Answers & Confidence
- Wire-compatible endpointA local server that accepts POST /v1/systemone and returns Choice, Score, or Noul answers in the documented JSON shape.Ecosystem
- Workflow evalsTypeSafe evalsTypeSafe's published evaluations of four automation workflows, comparing Jev and frontier LLMs as structured workflows versus single prompts.Evaluation & Training