adhd vs Praktika

The two products differ in kind, not degree. ADHD is an inference-time method with a reference implementation: N parallel branches per run, each generated under one of 15 cognitive frames (the paper names regulator, speedrunner, biology, and $0 budget), with no cross-branch context sharing so branches can't anchor on each other. Its generator system prompt forbids evaluation during branching, and generator and critic are mechanically separate LLM calls with opposing system prompts; a single critic pass then scores, clusters, and deepens top-K survivors. The branching driver is vantage-point reframing, not next-step variation. It ships as a skill and an npm package, integrates with the Claude and Codex Agent SDKs, and includes an eval suite plus a v0.1 preprint dated 2026-05-25. Praktika's feature set is entirely conversational: four named tutors for live voice and text practice, real-time corrections on pronunciation, grammar, and word choice, the tutor restating your sentence correctly and adding vocabulary, cross-session context memory, a personal study plan, grammar/vocabulary/listening/reading exercises, soft/balanced/strict feedback modes, a native-language interface, and free conversation on any topic. One has an API-level integration surface; the other has none listed beyond the iOS and Android apps.

ADHD costs nothing: pricing_type is free, and it's distributed openly on GitHub as a skill and an npm package with the eval suite and preprint included. Your real cost is compute and latency — running N branches plus a critic pass across the Claude or Codex Agent SDK is more expensive and slower per task than a single linear trace, which is why the method's own guidance steers latency- or cost-sensitive interactive agents away. Praktika is freemium: a free tier with conversation practice, and paid subscription tiers whose exact prices are not stated in the provided data. Its positioning is budget-conscious, including a September 2026 comparison of its pricing against four rival apps for Turkish learners on a traveler's budget and a 'small monthly budget' traveler use case; a September 2026 corporate-training piece also signals a business-buyer motion. Practical read: ADHD trade is tokens and wall-clock time against answer diversity; Praktika's trade is a recurring subscription against speaking reps you'd otherwise pay a human tutor for. Don't compare the numbers directly — one bill is API usage, the other a consumer app subscription.