Method and limitations
How does Talkallam evaluate spoken English?
A public explanation of what is measured, how CEFR is used, and what a score should not be taken to mean.
Use of the CEFR scale
Talkallam uses CEFR levels from Pre-A1 through C2 as a shared language for describing a learner's starting point and progress. The result is an in-product learning estimate, not Council of Europe endorsement or an official language certificate.
Post-session feedback dimensions
- Pronunciation
- Sound and rhythm clarity insofar as they affect intelligibility.
- Fluency
- Continuity, pauses, and the ability to complete an idea.
- Vocabulary
- Word choice, suitability, and range in the conversation context.
- Grammar
- Accuracy of structures that appeared in the learner's speech.
How a session becomes feedback
The system analyzes speech produced during the session, ties feedback to what was actually said, and presents actionable points instead of a context-free number. Sample length, microphone quality, noise, and connectivity can affect feedback accuracy.
Evaluation limits
Talkallam does not guarantee a specific learning outcome and does not replace an admissions, immigration, or professionally administered language examination. A score should be read with session examples and progress over time, not as a final judgment from one short sample.
Level estimates and feedback are learning tools, not a language certificate or official examination. Voice features for minors are not publicly available until parental-consent and safety gates are approved.
Voice and privacy
The Privacy Policy explains what is processed during a session and what may be retained. Voice features for minors are not publicly available before consent and safety approval.