Review queue
These repositories fell between the gates: Jev was confident enough not to drop them, but not confident enough to list them. The policy leaves them here rather than guessing. A human reading one repository for thirty seconds usually settles it.
If you can tell which way a row should go — or if you spot a wrong call anywhere in the index — open an issue with the Jev got it wrong template. Corrections become overrides in the data, and overrides feed the calibration set.
79 of 579 judged repositories are awaiting review.
Use Jev (TypeSafe's System One model) as a calibrated reranker: one call, up to 30 documents, a probability per document. Apache-2.0.
held because: category confidence 0.34 below 0.5
SDKs and ClientsRankingPythongenuine0.98Full judgmentNo description provided.
held because: category confidence 0.46 below 0.5
ApplicationsControl LoopTypeScriptgenuine0.96Full judgmentNo description provided.
held because: category confidence 0.45 below 0.5
Agent and Dev ToolingControl LoopJavaScriptgenuine0.96Full judgmentA playground for experiments around Jev, TypeSafe's System One model.
held because: category confidence 0.49 below 0.5
LearningFeature ScoringRustgenuine0.96Full judgmentSidecar: grade a cloudforge student writeup with TypeSafe Jev. Not part of the OSS product.
held because: category confidence 0.47 below 0.5
ApplicationsFeature ScoringPythongenuine0.96Full judgmenthl/jen
★ 1No description provided.
held because: category confidence 0.44 below 0.5
Agent and Dev ToolingVerificationGogenuine0.95Full judgmentmacOS computer use driven by Jev (TypeSafe System One) as the decision maker
held because: category confidence 0.44 below 0.5
ApplicationsControl LoopGogenuine0.95Full judgmentNo description provided.
held because: category confidence 0.24 below 0.5
OtherInfraJavaScriptgenuine0.95Full judgmentBounded exploratory browser testing with Jev, deterministic assertions, and replayable evidence.
held because: category confidence 0.39 below 0.5
SDKs and ClientsControl LoopTypeScriptgenuine0.94Full judgmentI tortured Jev into being a RISC-V CPU.
held because: category confidence 0.49 below 0.5
Games and SimulationControl LoopPythongenuine0.94Full judgmentPut your code comments on trial. Powered by Jev.
held because: category confidence 0.49 below 0.5
SDKs and ClientsFeature ScoringTypeScriptgenuine0.94Full judgmentNo description provided.
held because: category confidence 0.43 below 0.5
Research and EvalsControl LoopTypeScriptgenuine0.93Full judgmentNo description provided.
held because: category confidence 0.43 below 0.5
Agent and Dev ToolingRankingPythongenuine0.92Full judgmentJev (TypeSafe) 性能評価プロジェクト — 日本郵便 KEN_ALL をマスタに、AI SDK 経由の Jev が住所のあいまい一致にどこまで使えるかを検証
held because: category confidence 0.43 below 0.5
ApplicationsExtractionTypeScriptgenuine0.90Full judgmentPre-install security gate for npm lifecycle scripts using TypeSafe System One.
held because: category confidence 0.40 below 0.5
SDKs and ClientsVerificationJavaScriptgenuine0.90Full judgmentNo description provided.
held because: category confidence 0.49 below 0.5
ApplicationsRoutingPythongenuine0.89Full judgmentAn observable raw-character chat experiment powered entirely by TypeSafe Jev Choice
held because: category confidence 0.41 below 0.5
LearningControl LoopJavaScriptgenuine0.88Full judgmentExperimental multi-horizon BTC signal generator using TypeSafe Jev probabilities and Binance market data.
held because: category confidence 0.43 below 0.5
Research and EvalsControl LoopTypeScriptgenuine0.87Full judgmentReliability and evaluation harness for financial agent workflows: agent harness + evaluation harness, release gates, evidence packs, DSPy programs vs typed tool-use (Jev). Reference prototype, synthetic and public benchmark data.
held because: category confidence 0.44 below 0.5
Agent and Dev ToolingRoutingPythongenuine0.86Full judgmentjexp/neo4jev
★ 2Typesafe.ai System One Model Jev navigating a Neo4j graph by using a classifier over neighbouring relationships
held because: category confidence 0.42 below 0.5
LearningControl LoopJupyter Notebookgenuine0.85Full judgmentNo description provided.
held because: category confidence 0.30 below 0.5
LearningRoutingTypeScriptgenuine0.84Full judgmentHigh-speed recursive AI Elo tournament engine powered by Jev and Swiss matchmaking
held because: category confidence 0.43 below 0.5
ApplicationsRankingPythongenuine0.83Full judgmentHyprland/NixOS fork of typesafe-computer-use — grim + hyprctl + ydotool, TypeSafe Jev decisions
held because: category confidence 0.31 below 0.5
IntegrationsControl LoopPythongenuine0.83Full judgmentOne grounded Jev/Playwright core: typed SDK, persistent CLI, and MCP server with native browser operations and deterministic assertions.
held because: category confidence 0.36 below 0.5
Agent and Dev ToolingControl LoopTypeScriptgenuine0.82Full judgmentClassify Git commit diffs and messages with Jev. Bug fixes, security fixes/CWEs, and change types.
held because: category confidence 0.49 below 0.5
ApplicationsExtractionRustgenuine0.81Full judgmentACME live support-call scoring demo with TypeSafe AI, Effect, SQLite, React, Vite, and Turborepo
held because: category confidence 0.44 below 0.5
ApplicationsFeature ScoringTypeScriptgenuine0.79Full judgmentNo description provided.
held because: category confidence 0.47 below 0.5
Agent and Dev ToolingRoutingTypeScriptgenuine0.78Full judgmentOpen-source Jev log triage for OpenTelemetry. Score the signal before expensive LLM analysis.
held because: category confidence 0.39 below 0.5
ApplicationsFeature ScoringTypeScriptgenuine0.72Full judgmentAutomated database migration safety reviewer powered by TypeSafe AI (Jev System One model)
held because: category confidence 0.44 below 0.5
Agent and Dev ToolingVerificationTypeScriptgenuine0.61Full judgmentmy stuff for pi
held because: genuine 0.59 below 0.6
Agent and Dev ToolingVerificationTypeScriptgenuine0.59Full judgmentAutonomous PlayStation 2 AI Agent with real-time visual telemetry HUD powered by TypeSafe Jev System One
held because: genuine 0.59 below 0.6
Games and SimulationControl LoopPythongenuine0.59Full judgmentSelf-improving context compiler for CoreWeave Hacks: Agent Loops 2026
held because: genuine 0.58 below 0.6
Agent and Dev ToolingUnclearPythongenuine0.58Full judgmentVibe coding experiment
held because: genuine 0.58 below 0.6
Games and SimulationControl LoopPythongenuine0.58Full judgmentSystem-architecture skill for TypeSafe AI Jev/System One — find fuzzy semantic judgment and turn it into small Choice/Score/Noul primitives.
held because: genuine 0.58 below 0.6
Agent and Dev ToolingInfragenuine0.58Full judgmentopenvons (open-Jev): 有限選択肢に確率で答える判断層 — テキスト / 画像 / 日本語音声コマンド
held because: genuine 0.56 below 0.6
Research and EvalsControl LoopPythongenuine0.56Full judgmentSO-101 robot-arm agent workbench: Bun/Effect coordinator, React workbench, Python LeRobot motor owner
held because: genuine 0.55 below 0.6
Games and SimulationControl LoopTypeScriptgenuine0.55Full judgmentBenchmarking TypeSafe's Jev decision model as a cost-efficient LLM router on RouterArena
held because: genuine 0.55 below 0.6
Research and EvalsRoutinggenuine0.55Full judgmentA simple way to deal with streaming text, tools, images from LLMs.
held because: genuine 0.54 below 0.6
SDKs and ClientsInfraGogenuine0.54Full judgmentAgent tool/MCP call gate — allow / ask_human / deny via TypeSafe Jev
held because: genuine 0.54 below 0.6
Agent and Dev ToolingVerificationTypeScriptgenuine0.54Full judgmentshiro-0x/hersona
★ 51346 reusable character attributes for AI agent personas — compose, measure, and port system-prompt personas. Build once. Keep personality everywhere.
held because: genuine 0.52 below 0.6
Agent and Dev ToolingRoutingPythongenuine0.52Full judgmentEvidence-aware local RAG and HandoffProof: controlled causal testing for operational handovers, with optional TypeSafe/Jev evidence governance.
held because: genuine 0.51 below 0.6
ApplicationsVerificationPythongenuine0.51Full judgmentNo description provided.
held because: genuine 0.51 below 0.6
Agent and Dev ToolingUnclearZiggenuine0.51Full judgmentJev (TypeSafe System One) × ASReview SYNERGY abstract screening demo — Choice/Noul vs gold labels
held because: genuine 0.51 below 0.6
Research and EvalsRankinggenuine0.51Full judgmentAn experimental protocol for evidence-aware agent handoffs, bounded worker continuation, and TypeSafe/Jev-assisted review, with reproducible evaluation.
held because: genuine 0.50 below 0.6
Agent and Dev ToolingVerificationPythongenuine0.50Full judgmentTypeSafe Jev research + 5 product specs
held because: genuine 0.50 below 0.6
OtherUncleargenuine0.50Full judgmentAn evaluation of typesafe AI chess. As it turns out, the AI isn't doing really well even though chess is not a particularly open-ended game. Still, it's only a prototype and this probably wasn't optimzied for games.
held because: genuine 0.49 below 0.6
Research and EvalsUnclearPythongenuine0.49Full judgmentWorkload-aware Pareto frontiers and dynamic model selection for agentic systems
held because: genuine 0.47 below 0.6
Research and EvalsRoutingPythongenuine0.47Full judgmentvercel-labs/ai-cli
★ 705Generate anything from your terminal
held because: genuine 0.46 below 0.6
SDKs and ClientsRoutingTypeScriptgenuine0.46Full judgmentJev-style parallel constrained decisions for any MLX model on Apple Silicon. Typed, schema-valid JSON in one forward pass.
held because: genuine 0.46 below 0.6
Research and EvalsExtractionPythongenuine0.46Full judgmentJudge agent steps — ok / retry / escalate / stop via TypeSafe Jev
held because: genuine 0.46 below 0.6
Agent and Dev ToolingUnclearTypeScriptgenuine0.46Full judgmentPulseLane — clinic triage decisions via TypeSafe Jev
held because: genuine 0.46 below 0.6
ApplicationsUnclearTypeScriptgenuine0.46Full judgmentNo description provided.
held because: genuine 0.45 below 0.6
Agent and Dev ToolingInfraPythongenuine0.45Full judgmentA typed decision-routing prototype for turning medical learning material into deterministic study actions, designed to evaluate TypeSafe Jev.
held because: genuine 0.44 below 0.6
ApplicationsRoutinggenuine0.44Full judgmentNo description provided.
held because: genuine 0.43 below 0.6
LearningUnclearPythongenuine0.43Full judgmentAgent budget brake — continue / downgrade_model / stop via TypeSafe Jev
held because: genuine 0.42 below 0.6
ApplicationsUnclearTypeScriptgenuine0.42Full judgmentSelf-correcting audiobook TTS generator (CoreWeave Hacks) — TypeSafe tagging + Google TTS/STT + ARIA/Weave eval-correction loop
held because: genuine 0.42 below 0.6
ApplicationsControl LoopPythongenuine0.42Full judgmentLocal bilingual probability decisions from context, questions, and candidate answers. Independent research preview inspired by TypeSafe Jev.
held because: genuine 0.41 below 0.6
Research and EvalsInfraPythongenuine0.41Full judgmentu007/ocode
★ 1No description provided.
held because: genuine 0.41 below 0.6
Agent and Dev ToolingVerificationGogenuine0.41Full judgmentTyped judgment layer for coding agents — gates from PRD to ship. Jev-ready, provider-agnostic.
held because: genuine 0.40 below 0.6
Agent and Dev ToolingVerificationTypeScriptgenuine0.40Full judgmentMatch user goals to MCP catalog (two-stage) via TypeSafe Jev
held because: genuine 0.39 below 0.6
ApplicationsRankingTypeScriptgenuine0.39Full judgmentTypeSafe Jev MCP decision layer for coding agents and CI
held because: genuine 0.39 below 0.6
Agent and Dev ToolingUncleargenuine0.39Full judgmentNo description provided.
held because: genuine 0.39 below 0.6
LearningUnclearTypeScriptgenuine0.39Full judgmentThe open-source email platform — self-host on your own AWS SES, or use the cloud. Resend-compatible API.
held because: genuine 0.38 below 0.6
ApplicationsFeature ScoringTypeScriptgenuine0.38Full judgmentFilters an agent's memories to fit a token budget. Local, HTTP, MCP, Docker.
held because: genuine 0.38 below 0.6
Agent and Dev ToolingRankingPythongenuine0.38Full judgmentRoute tasks to research/code/browser/support/writer agents via TypeSafe Jev
held because: genuine 0.38 below 0.6
Agent and Dev ToolingRoutingTypeScriptgenuine0.38Full judgmentTypesafe AI SDK in Elixir using Req
held because: genuine 0.38 below 0.6
SDKs and ClientsInfragenuine0.38Full judgment3D chess powered by TypeSafe AI (Jev). AI vs AI by default, or play either side. Multiple difficulty levels.
held because: genuine 0.37 below 0.6
Games and SimulationUnclearTypeScriptgenuine0.37Full judgmentA polished OpenAI + TypeSafe Jev terminal interface for answers with transparent decision reports
held because: genuine 0.37 below 0.6
SDKs and ClientsUncleargenuine0.37Full judgmentJev-style parallel constrained decisions for any MLX model on Apple Silicon. Typed, schema-valid JSON in one forward pass.
held because: genuine 0.36 below 0.6
Research and EvalsExtractionPythongenuine0.36Full judgmenttypesafe.ai model jev finance benchmark
held because: genuine 0.36 below 0.6
Research and EvalsUncleargenuine0.36Full judgmentU-NSGA-III multi-objective evolutionary optimization for .NET (Seada & Deb) — ZDT/DTLZ, IGD oracle vs pymoo
held because: genuine 0.35 below 0.6
OtherRankingC#genuine0.35Full judgmentLaneBreak — support ticket priority+routing via TypeSafe Jev
held because: genuine 0.35 below 0.6
ApplicationsUnclearTypeScriptgenuine0.35Full judgmentTyped-decision benchmark from PadFlow (land development SaaS): schemas, anonymized labeled rows, and a runner for confidence-calibrated models like TypeSafe Jev.
held because: genuine 0.33 below 0.6
Research and EvalsRoutingPythongenuine0.33Full judgmentAutonomous prospecting for freelancers.
held because: genuine 0.33 below 0.6
ApplicationsUnclearHTMLgenuine0.33Full judgmentHireSignal — resume first-pass fit+interview via TypeSafe Jev
held because: genuine 0.33 below 0.6
ApplicationsFeature ScoringTypeScriptgenuine0.33Full judgmentNPCs of River Oaks Houston, Texas using Jev to power NPCs
held because: genuine 0.33 below 0.6
Games and SimulationUncleargenuine0.33Full judgmentA Jev-inspired decision interface for existing LLMs. Explicit choices, scores, calibration, and review thresholds.
held because: genuine 0.32 below 0.6
IntegrationsRoutingHTMLgenuine0.32Full judgmentOffline evaluation for document extraction
held because: genuine 0.32 below 0.6
OtherVerificationPythongenuine0.32Full judgmentA test project based on Jev AI, the goal is to build a search function for a blog/article website that has 100s of articles to search from, So the user can actually use the search as chat to question anything and find related answers/articles
held because: genuine 0.31 below 0.6
ApplicationsUnclearJavaScriptgenuine0.31Full judgment