Mind Forge is a small, independent library of 24 free cognitive exercises. Each one is built around a real task used in psychology research — not a repackaged mobile game. We tell you honestly what each tool measures, who it suits, and where the science gets wobbly.
Each card links to a standalone page. Pick one that matches what you want to exercise today.
Audio + visual streams, with auto-leveling that tracks your real threshold (not just session best).
Unique: per-modal hit/false-alarm breakdown chart.
Color-word interference test with reaction-time histograms and a fatigue detector.
Unique: detects RT drift across the session.
Find numbers 1–25 in order, with per-cell time heatmap so you can see your scan pattern.
Unique: visual scan-path heatmap.
Reproduce growing sequences on a 3×3 board; reverse-mode toggles visuospatial span.
Unique: forward & reverse span tracking.
Judge if rotated 3D shapes match; includes angle-difficulty curve estimation.
Unique: fits your personal rotation-rate slope.
Tap for "go" stimuli, hold for "no-go". Audio + visual cues measure cross-modal control.
Unique: commission vs omission error breakdown.
Auditory digit sequences with backward recall, audio-rate control.
Unique: adjustable digit presentation rate.
Connect numbers, then alternate with letters. Ratio reveals set-shifting cost.
Unique: per-segment timing & error map.
Three (and six bonus) classic trick problems. We log intuitive vs reflective answers.
Unique: tracks which impulse you fought.
Find a target among distractors. Reaction time vs set size gives your search slope.
Unique: feature vs conjunction search.
Two scenes alternate with a gray blank. Spot the change before your attention slips.
Unique: incremental scene-degradation mode.
Quick arithmetic with adaptive difficulty based on rolling accuracy and RT.
Unique: per-operation accuracy breakdown.
Memorize lists A, B, C in sequence; release from proactive interference is measured.
Unique: per-list decay curve plotting.
Discover your own chunks in unfamiliar digit strings; teaches chunking-by-pattern.
Unique: shows how you spontaneously chunk.
Estimate quantities (count, distance, weight). Get a calibration score, not just "right".
Unique: confidence calibration curves.
Embedded target detection: press a key when X appears during an unrelated task.
Unique: event-based vs time-based cues.
Two simultaneous choice tasks. Compare dual-task cost to single-task baseline.
Unique: cost = dual − single, plotted per block.
Watch colored shapes follow hidden rules; predict the next. Rule-complexity grows.
Unique: rule-difficulty curve fitting.
Sort cards; rule changes without warning. Perseveration rate is the headline metric.
Unique: trails-to-first-correct + perseveration.
Recognize abstract faces after a delay; tests associative visual memory.
Unique: same/different rather than recall.
Place items along a familiar route, then recall. Built-in loci hints for beginners.
Unique: route rehearsal feedback.
Watch for a specific cue-then-target sequence; rate of misses reveals vigilance drift.
Unique: time-on-task drift plot.
Read sentences and remember final words; classic complex-span working-memory test.
Unique: partial-credit scoring & serial-position curve.
Press the matching arrow key; Hick's law check with per-choice histograms.
Unique: Hick-rate constant estimation.
These are not apps that promise to "raise your IQ in 7 days." They're reproducible cognitive tasks — useful if you treat them like a researcher would.
Memory? Attention? Reasoning? Each card is tagged so you can target a specific cognitive function rather than scrolling randomly.
Your first attempts are mostly learning the interface. We recommend 3 sessions on different days before reading meaning into a number.
All your data lives in localStorage on your device. Look at trajectories across days/weeks, not yesterday's number.
Honest disclaimer
Improving on a specific cognitive task doesn't automatically mean your "brain" got better in everyday life. The transfer question is the most-debated topic in this field. We tell you where the evidence is thin in every tool's limitations section.
Mind Forge is an independent, single-developer project. It's built by people who are skeptical of the brain-training industry but genuinely curious about what happens when you put a real cognitive task in front of someone for ten minutes a day.
Every tool here is modeled on a published experimental task. We link the canonical references inside each tool's notes section. The design choices — color palette, font size, default difficulty — were picked to not interfere with what the task actually measures.
Like most small independent sites, we display advertising (Google AdSense) so the project can stay free. You can read exactly how that works in our privacy section below.
Honest answer: we don't know for sure. The transfer question — does practicing a narrow cognitive task improve everyday cognition — is the most-contested topic in this field. Recent large studies (like the BBC-funded "Brain Test Britain" study with over 100,000 participants) found no evidence of broad transfer. So treat these as enjoyable, repeatable practice on the task itself, not a guarantee of real-world gains.
No. Your scores, settings, and history are stored only in your browser's local storage. We never upload personal training data. Ads may set cookies per Google's standard AdSense practices — that's covered in the privacy section.
Most tools are designed for 10–20 minutes per session, 2–4 times per week. Going longer doesn't help much — fatigue effects kick in and the data becomes less useful.
They target different cognitive functions. Dual N-Back and Corsi Blocks are notoriously hard for most adults; Mental Rotation and Math Reflex feel easier because they tap into well-practiced skills. That's expected — pick what you find engaging, not what sounds impressive.
These tools are recreational/educational. They are not validated clinical assessments, and they have not undergone the psychometric certification required for diagnostic use. For research-grade instruments, use validated software such as Inquisit, PsychoPy, or PARiConnect.
Yes. We add a new tool every few months based on user suggestions and published research. The most-requested tools tend to be ones that have a clean published protocol.