Real-Time Teacher-Led Adaptive Learning
Live context stays on the device. When a session closes coherently, traceable evidence can enter durable memory and later return as a small, scoped packet for a teacher decision.
What continues after the call
The original lesson moment, a teacher-confirmed correction, later recall and the next decision stay connected. Each step belongs to the same human-led learning path.
The teacher notices and decides. LingoSide preserves where the evidence came from, brings back the relevant part and leaves room to say there is not enough evidence.
See the product layersDuring the lesson
Observe the lesson
Keep source text, translation and contextual reply suggestions in a sidecar that takes one glance, so a meaningful phrase, correction or difficulty can be noticed while the conversation keeps going. Earlier moments from the same session inform what is suggested.
After the lesson
Keep what happened
Preserve the session as a finalized record: saved phrases, teacher-confirmed corrections and evidence-linked moments. Every item keeps a pointer to the exact utterance it came from, so it can be checked against the lesson instead of taken on trust.
Between lessons
Practise it later
Ask the learner to retrieve first, then reveal the source moment and its context. Selected material comes back at useful intervals and in the situation where it first mattered, rather than as another generic list of exercises.
At every teaching proposal
Teacher decides
Show the teaching proposal together with the lesson moment behind it, so the teacher can approve, edit or reject it. That decision, not the model’s confidence, is what the rest of the loop builds on.
Before the next lesson
The next lesson adapts
Bring the right earlier material back into reach: due reviews, saved phrases, open corrections and possible patterns, each one leading to the evidence behind it. LingoSide is building this stage so that earlier decisions support better choices for learner and teacher, without turning the lesson into a dashboard or an automated course.
Evidence-based memory
LingoSide does not need to remember everything. It preserves useful lesson evidence and the teacher’s correction, and it is building the memory of what comes next: the learner’s later recall and what happens when the language returns.
A phrase, correction or difficulty remains linked to the source utterance.
Teacher confirmation, edits and rejection become part of the record.
The learner tries to retrieve before seeing the answer and context again.
The loop is designed to distinguish independent recall, transfer, prompted use and no attempt.
An observation can be checked against the lesson moment that supports it.
When a teacher corrects or removes an item, later views are designed to use the corrected record.
A hypothesis is not presented as a fact about the learner.
Teacher authority
LingoSide can surface a traceable moment and a possible next move. The teacher reads the learner, chooses the timing and decides what happens in the lesson.
Observation
The lesson evidence suggests a retrieval problem, not proof of a broad weakness.
Possible strategy
Bring the phrase back through active recall in a related conversation.
Under the hood
During a lesson, LingoSide serves the active working set from client RAM. With session archiving enabled (the default), it checkpoints finalised content locally. After the session closes coherently, its Evidence is intended to enter the backend index for durable, correctable recall across lessons.
For everyone
Think of it as a librarian with a strict access card: search only the right learner and course, combine exact-word and meaning matches, then hand the AI a few traceable excerpts—not the whole history.
Established terms, one architecture
RAM + default-on local checkpoints + backend memoryRAG (retrieval-augmented generation) means that before AI helps, a retrieval layer looks for relevant Evidence instead of pouring the whole transcript history into the prompt. The terms below name separate parts of the architecture; together they describe LingoSide’s design, not a new industry standard.
Finalised moments keep their speaker, language, time and source. Across lessons, a raw transcript or AI hypothesis is not treated as a fact: teacher confirmation, uncertainty and correction status stay explicit.
The server limits candidates to the permitted learner and course before Evidence is ordered. Structured fields remove invalid or superseded items; full-text search (FTS) and embeddings then find exact and semantic matches.
Time, review state, teacher-confirmed corrections and the current lesson goal can change the order of Evidence candidates. The consumer receives only top-k Evidence with provenance and why it matched—or no result when confidence is too low.
Access is checked before retrieval, so only authorised people can read learner-and-course memory.
Corrections and invalidations affect recall, so stale evidence does not quietly return.
Corpus size is not prompt size: AI receives a small, traceable packet instead of the whole history.
In deployment terms, the edge–cloud split has two client-side layers: RAM serves latency-sensitive working state, while default-on local session checkpoints protect finalised lesson content during the session. Only after the session closes coherently is cross-lesson Evidence intended to enter the server-side index. That topology is separate from hybrid retrieval, which describes how backend Evidence is found.
Structured filters, FTS and embeddings generate Evidence candidates; temporal and lesson-goal signals order them by relevance. Only the best small packet reaches the consumer. A larger history does not require GraphRAG or a longer context window by default.
Every retained item keeps its source, date, language, status and teacher decision. A correction creates a new version; invalidated or superseded evidence leaves active recall without erasing the audit trail.
Search is scoped on the server to the requester, learner and course before Evidence is ordered by relevance. Without the required grant, a requester sees nothing, however similar the text may be.
Session boundaries
The learner sidecar holds no provider key of its own. It sends one speech stream for live recognition, from the source selected for the lesson: the learner microphone or system audio from the lesson call. The teacher sidecar reads the shared lesson state and opens no audio connection.
Two connected sidecars. One speech stream from the learner sidecar: the source selected for the lesson, its microphone or system audio from the lesson call. It goes directly to the STT provider for live recognition. The teacher sidecar reads the shared lesson state, transcript and teaching support; it sends no audio.
The Windows package is built without a provider key inside it, so it cannot connect to anything on its own. It has to ask the backend first, every single time.
Each connection attempt gets its own pass that expires about a minute later. Nothing long-lived sits on the computer waiting to be copied or reused.
The backend opens the session, issues the pass and records what happened. The sound itself goes directly from your machine to the recognition service, so there is no shared middle stop to wait behind.
For a teacher this means one practical thing: putting a lesson through LingoSide does not drop it into a shared pipe, and it does not leave a reusable key sitting on the computer afterwards.
Where it runs
A lesson behaves the same wherever it runs, because this is not four products pretending to be one. The sidecar is a single application compiled for each system.
The desktop sidecar, delivered as a signed package that updates itself in the background.
The same desktop sidecar on Apple hardware.
A phone beside the lesson rather than a second screen to manage.
The same phone-sized sidecar on iPhone and iPad.
Flutter is Google’s open-source toolkit for building one application in Dart and compiling it into a real native app for each operating system — not a website in a wrapper. LingoSide uses it for the whole sidecar, so the lesson logic, the interface and the backend contracts live in one codebase instead of four. That is why adding a platform is mostly packaging, permissions and store rules rather than rebuilding the product.
The learner sidecar sends one speech stream for recognition, from the source selected for the lesson: its microphone or system audio from the lesson call. The teacher sidecar reads the shared lesson state and sends no microphone or system audio.
Technology behind LingoSide
Data architecture
Events move through the system, but their meaning stays owned. Stream processing, the learning-data warehouse, AI diagnostics and product analytics remain separate by design.
Domain event path
In the target architecture, the client event stream enters Proton, where streaming SQL shapes live events and sends durable batches to ClickHouse.
Real-time event processing
Proton OSS is Timeplus’s Apache 2.0 open-source streaming SQL engine: a single C++ binary with no JVM or ZooKeeper. The target architecture assigns it the client event stream and streaming SQL, with durable batches routed to ClickHouse.
Evidence for our own benchmark—not a capacity promise.
Why a separate layer
The live stream can absorb pressure and be replayed without making the domain depend on a stream engine API.
Domain warehouse: RAW + Serving
The target architecture assigns ClickHouse the durable learning-data warehouse—not the live stream, AI trace store or product telemetry.
RAW · evidence history
Append-oriented events and educational evidence. New interpretation never overwrites the historical record.
Serving / Derived · usable state
Rebuildable projections, aggregates and current learner state used by the backend.
Same RAW + same derivation version → same semantic state.
Why a separate layer
Evidence stays auditable; current state can change and still be reproduced from history.
AI execution observability
The backend’s metadata-only trace layer links prompt versions, models, retrieval stages, latency and cost. Typed evaluations can join this layer when they exist.
Why a separate layer
Engineers need the execution waterfall without copying lesson content into diagnostics.
Product analytics and experiments
Product behaviour belongs here only through explicit consent and a closed allowlist: flows, feature use, feature flags and experiment outcomes.
Why a separate layer
Product behaviour can be measured without turning raw educational data into analytics.
Teacher Action Loop projection
A read-only view correlates domain evidence with AI execution metadata. Stable references cross the boundary; the underlying records stay with the system that owns their meaning.
Domain path · ClickHouse
ClickHouse contributes event and evidence IDs, causation, the teacher action, the executed intervention and the next learner attempt.
AI path · Langfuse
Langfuse contributes the AI trace ID, prompt and config digests, model, retrieval stages, latency and cost—metadata only.
Correlated result
The read-only Teacher Action Loop projection joins both paths through stable references. It owns no canonical data; a missing critical edge returns INCOMPLETE TRAJECTORY instead of a guess.
Typed importer boundary
The import contract gives a typed projection importer one job: read authorised domain rows from ClickHouse and metadata-only trace records through Langfuse’s supported API or export boundary, then map stable IDs into the read-only trajectory shape. It does not depend on Langfuse’s internal ClickHouse tables or copy lesson content.
Supporting stack
Speech recognition and AI services connect to LingoSide through specialized, replaceable adapters. The learning layer depends on stable internal contracts rather than any single vendor’s protocol. For speech recognition, that contract includes timestamped segments, speaker attribution, confidence and language identification. Speaker attribution separates voices; it does not name them. Which voice is the learner and which is the teacher is a separate decision, confirmed in the lesson and correctable there, and it stays unassigned until someone confirms it.
Audio is sent directly from the learner sidecar to the STT provider; it is not relayed through the LingoSide backend. To evaluate ASR integrations, LingoSide is building a separate, versioned test corpus and a streaming harness that measures accuracy, code-switching, speaker attribution, partial-result revisions, reconnect behavior and latency. It also replays saved structured session events—not production audio—through the downstream learning pipeline.
This architecture allows providers to be evaluated or replaced as the technology improves without rebuilding the lesson experience.
ASR evaluation harness
LingoSide is building a provider-neutral ASR Evaluation Harness to judge speech systems against evidence, not vendor claims.
Versioned fixtures and lesson scenarios will exercise replay and live streaming, with evidence separated by language, condition and failure mode.
Transcript quality
Versioned reference transcripts will make accuracy, code-switching and critical lesson terms comparable across repeatable language slices.
Streaming quality
Partial revisions, time to stable words, end-to-final latency and reconnect recovery will be evaluated as a sequence—not collapsed into the final sentence.
Attribution quality
Speaker attribution will be checked against labelled scenarios independently of transcript accuracy. A correct sentence assigned to the wrong speaker still fails the scenario.
The evidence standard
No quality claim without a versioned fixture, an expected result and a reproducible run.
Harness engineering
LingoSide uses a versioned quality contract to define risk and required proof. Harness engineering turns that contract into a repeatable question for each learning, AI and speech path: what evidence must pass before a release?
The operating rule
Claims and changes must be testable against evidence.
Input
A candidate enters as something to test, not something to trust.
Exercise
Contracts · replay · evaluations · fault cases
Verify
Tests · privacy · regressions · required evidence
Review
After repository tests, Code Review examines the implementation; Decision Review checks alignment with product decisions. Open findings return the change for revision.
Decide
One method, four boundaries
Lesson flow, speech, teacher AI and later recall are different surfaces. They share the same demand: a scenario, an expected outcome and evidence that can fail.
01 · Lesson path
The executable lesson contract covers a 60-minute lifecycle, six EN/PL/RU directions, archive reopen and four controlled failures. That makes lifecycle, language coverage, recovery and failure handling reviewable as one contract.
No LingoSide audio recording · The harness replays a synthetic event scenario, not a sound file.
02 · Speech path
A defensible ASR quality gate requires a versioned corpus and streaming measurements sliced by accuracy, code-switching, speaker attribution, partial-result churn, reconnects and latency. Missing coverage is a gap, never proof.
03 · Teacher AI
Provider-neutral policy checks block invalid suggestion shapes. Beyond shape validation, a defensible AI evaluation requires evidence-linked teacher scenarios and keeps the final decision with the teacher.
04 · Learning path
The learning data model preserves review attempts and outcomes. A defensible learning evaluation compares those records with delayed recall instead of treating engagement as learning.
How replay works
Replay is not ‘press play.’ LingoSide does not create or save an audio recording. The contract reruns a versioned synthetic lesson scenario and compares the structured result with explicit expectations.
Synthetic transcript events carry text, translations and timestamps; the scenario also declares language pairs and controlled failures.
The harness injects the same synthetic events into the production session controller and advances a virtual lesson clock.
It checks the expected outcomes, opens the resulting structured session in a fresh archive store and verifies the archive UI and failure paths.
A real lesson starts differently: audio goes directly to the STT provider for live recognition, while LingoSide archives structured outputs—not a recording. Speech accuracy is a separate gate: it requires an explicit, versioned audio corpus.
LingoSide Research Lab
We already know how to run a live lesson. We are researching how to make its intelligence measurable, replaceable, resilient, and trustworthy under real classroom conditions.
Research experiment 01
While building LingoSide, problems emerge that current tools do not solve well enough. Research Lab documents these experiments. Not all of them will become part of the product or the tools LingoSide uses.
We are testing whether one admission protocol can evaluate many typed evidence contracts without flattening them.
Executable harnesses, canaries and eval runners keep their own typed evidence models. The proposed shared envelope is designed to let an action-specific gate verify what ran, what it checked and whether the evidence is sufficient—without forcing every result into one model.
Scope: immutable artifacts from executable harnesses, canaries and eval runners. Telemetry, human decisions and market research stay outside. Evidence Contract, Evidence Admission Gate and Verification Decision Record are LingoSide working names—not industry terms, products or proposed standards.
Evidence boundary
Executable producers already emit incompatible schema_version: 1 formats. Each payload therefore stays independently typed and validated; telemetry, human decisions and market research keep their own models outside this mechanism.
This mechanism is valid only while adding a producer leaves the envelope unchanged; consumers do not privately parse freshness, policy or producer results; domain fields stay out of the envelope; payload data stays typed; migrations preserve existing results; and shared parsing removes more duplication than it adds. Gate, registry and policy changes cannot approve themselves; they require a trusted base policy or an explicit owner Human Gate.
Prior art, not a novelty claim
in-toto separates immutable subjects and predicate types from typed payloads, and catalogs separate verification-result predicates. OPA separates policy decisions from enforcement. A SLSA example records subject, verifier, policy digest and verification result. LingoSide is testing a local fit, not claiming an invention.
Producer-specific typed validation; a minimal envelope with no producer-assigned authority; action-specific admission under trusted policy; a separate verification decision record; and fail-closed unknown schemas or action policies.
Runtime telemetry and state, surveys, interviews, market research, one global evidence score and one shared domain payload remain outside. Human Gate stays separate and references digests. Cryptographic runner identity, DSSE and Sigstore are not part of the first slice. This is not a product, standard, SDK, plugin system, open-source project or SaaS.
Evidence Control Plane · read-only admission projection
Evidence Control Plane is a read-only mockup that keeps the producer result, typed validation, action-specific gate decision and human decision separate. It cannot run a producer or policy, grant authority, or create any decision record.
The console shows no lesson text, prompts, model output, audio or secrets. Without authenticated attestation identity it can check structure, digests and revision linkage—not cryptographically prove who ran the producer.
All five live-model cases matched the expected response mode and policy outcome. That describes what the canary observed; the producer grants no permission to merge or release.
Each producer payload has its own identifier and typed validator. This canary keeps its domain fields; the envelope only identifies and hashes them.
These provider-specific measurements stay in the canary payload. They are not fields in the shared envelope.
Move from a scenario outcome to what was expected, what was observed, the source artifact and the reproduction command.
mode=short_hint · variants=1PASSmode=silence · variants=0PASSmode=direct_form · variants=1PASSmode=direct_answer · variants=1PASSmode=examples · variants=2PASSThe gate—not the canary—would combine the requested action, validated envelope, exact subject and run context, trusted policy and producer result.
No release_candidate policy exists in this mockup, so the gate fails closed. A future comparison policy could classify the same report as SCREENING without granting release authority.
The console shows no lesson text, prompts, model output, audio or secrets. Without authenticated attestation identity it can check structure, digests and revision linkage—not cryptographically prove who ran the producer.
The producer result, typed validation, consumer action and trusted policy remain distinct inputs to a separate admission decision.
Whether AI chooses the expected response mode and whether the result passes the shared teaching policy.
An early regression signal, not a statistical quality measure or evidence of system behavior throughout a full production lesson. The producer result does not authorize merge or release; only an action-specific admission policy could.
The mockup keeps the typed payload, minimal envelope and gate-produced verification decision distinct. A human decision remains a separate object that references digests.
Research protocol candidate 02 · Teacher Action Loop
A good teacher already adapts in the moment. LingoSide is testing whether that adaptation can become persistent: linked to the lesson evidence that triggered it, the teacher’s actual action, the learner’s next attempt, and a later probe that can confirm, correct, reject or expire the memory.
EVIDENCE LADDER · SEQUENCE MATTERS
02A cannot stand in for 02B. Neither can stand in for 02C.
Can the complete trajectory be observed, reviewed, corrected and reproduced?
Does evidence + proposed strategy change the teacher’s response versus the same evidence alone—within burden and disruption limits?
Does the preassigned support policy improve delayed production or transfer against a fair comparator?
DISCOVERY + CALIBRATION INPUTS · NOT TRIAL EVIDENCE
These studies help determine what to measure. They do not validate LingoSide.
Source evidence → system observation → diagnostic hypothesis or abstention. Eligibility closes before the suggestion or outcome.
Approve, edit, reject or do nothing—then keep intended action separate from actual action.
The next attempt and immediate outcome stay distinct from a click, a teacher preference or delayed learning.
A delayed probe leads to confirm, correct, reject or expire. Rejected or superseded hypotheses never return as current evidence.
WHY THE PROTOCOL NEEDS RESEARCH
15 studies · 827 learners
A classroom meta-analysis found durable benefits from oral feedback, with prompts outperforming simply supplying the corrected form.
Lyster & Saito · SSLASource checked ↗A decade of research
A research review identifies contingency, fading and transfer of responsibility as the core of teacher scaffolding. It also warns that effectiveness evidence is limited and the interaction is hard to measure.
van de Pol, Volman & Beishuizen · Educational Psychology ReviewSource checked ↗Whether the complete loop can be observed, reviewed, corrected and carried forward without reviving rejected interpretations.
Whether offering LingoSide support caused better teacher decisions, retention or transfer. Those require separate 02B and 02C comparisons of preassigned policies and delayed outcomes.
AI may propose a hypothesis and a possible action; the teacher owns the pedagogical decision. Evidence keeps its provenance and correction state, rejected evidence leaves active recall, weak evidence leads to abstention, and no learner score is produced.
Research experiment 03 · The Smallest Useful Memory
We are testing the smallest honest boundary for continuity between lessons: data volume, teacher review time and failure rate—measured separately.
Boundary frozen 11 Aug 2026dbc1b6c209ca8bab6a08f7d454920ff3d408ede6
PASS BOUNDARY · DECLARED BEFORE RESULTS
≤2 min teacher review / 60 min lesson
Same lesson · four memory policies
The full structured lesson history establishes the recall and unsupported-suggestion baselines. It is not assumed to be the desirable retention policy.
Automatic compression receives exactly the same byte ceiling as teacher-confirmed evidence, without teacher approval. This separates “less data” from “teacher confirmation”.
The teacher can confirm, correct or reject revisable evidence. Metadata, provenance, uncertainty and correction history all count inside the same 10% budget.
The consumer sees no state from lesson N. This shows which next-lesson outputs do not depend on retained context at all.
Motion explains the budget reduction only. It does not display a result.
H1 · MINIMAL MEMORY IS SUFFICIENT
H2 · TEACHER CONFIRMATION ADDS VALUE
Critical-context recall must improve by at least 10 percentage points, without worsening unsupported suggestions by more than 2 percentage points.
GROUND TRUTH · SEALED FIRST
It is sealed at the end of lesson N—before any memory output, held-out next-lesson scenario, AI suggestion or evaluation result.
recall = correctly used sealed entries / activated sealed entriesMINIMUM CONFIRMATORY SET
Opportunities are nested within lesson pairs and teachers. This minimum can falsify the protocol; it cannot establish population-level generalisation.
It calibrates the harness only and is excluded from every confirmatory numerator and denominator.
Whether teacher-declared continuity can fit within 10% retained state, preserve at least 90% of critical context and stay within the declared teacher-time cost in this test.
That less data caused learning, privacy risk disappeared, AI discovered what matters, or three teachers represent a population.
No result appears here until the closed run can report its numerator, denominator and uncertainty. Thresholds are not outcomes.
An engineering view
This is AI infrastructure for continuity of human-led learning: one selected audio source becomes one speech stream; provider adapters create structured events; provenance travels with evidence; automation stops before the teacher's decision.
one selected source · one speech stream
the provider protocol ends here
proof scope travels with the result
automation stops before judgment
Provider protocols end at adapters. Shared layers consume stable contracts.
Synthetic events replay the lesson contract without creating an audio recording.
Provenance and proof scope accompany every engineering conclusion.
Automation can propose. The teacher owns pedagogical judgment.
The architectural core is domain-independent: its invariants do not depend on language learning.
Language learning is the first domain, not a constraint of the architecture.