End-to-end correction workflows
Turn granular word- and phoneme-level feedback, advanced fluency analysis and prosody evidence into focused learner corrections.
SDK & API
Pronunciation assessment API & speech evaluation SDK for EdTech.
CHIVOX delivers industry-leading pronunciation assessment APIs and SDKs for language learning products and services. Engineered for precision speech scoring, our solutions provide granular word- and phoneme-level feedback alongside advanced fluency and prosody analysis.
iOS · Android · Web · ServerREST + WebSocket streamingEnterprise licensing via BDTrusted by 100+ customers worldwide
Word → passage
task coverage
EN + Mandarin
one scoring surface
JSON out
product-ready results
/best fit
Professional speech scoring, ready for EdTech at scale.
Power end-to-end pronunciation correction inside adaptive tutoring apps, language learning products and high-stakes testing environments. Robust enterprise deployment options help organizations ship professional scoring with minimal integration overhead.
Actionable JSON by default
Every assessment returns structured JSON your product can use for highlights, coaching, retries, analytics and adaptive next steps.
Trusted at enterprise scale
Flexible deployment options support organizations at scale. More than 100 customers worldwide rely on CHIVOX speech scoring.
/platforms
One engine across the surfaces you already ship.
Native SDKs for mobile and web, plus REST and WebSocket for backends — same scoring family whether the audio is streamed live or submitted after capture.
Mobile
iOS · Android
Native SDKs for apps that record inside a lesson, drill or exam flow.
Web
JS SDK
Browser capture for web tutors, classroom tools and online practice.
Server
REST · WebSocket
Backend scoring when audio is already uploaded or streamed from your stack.
Product surface
Your UX
Wire scores into meters, highlights, retries and coaching — not a fixed Chivox UI.
/task coverage
Assessment modes from a single sound to open speech.
English and Mandarin speaking tasks — from phonics and characters to paragraphs and interactive dialogue — through one SDK/API surface.
English · 12 kernels
- Natural phonicsen.nsp.score
- Word assessmenten.word.score
- Word correctionen.word.pron
- Vocabulary seten.vocabs.pron
- Sentence assessmenten.sent.score
- Sentence correctionen.sent.pron
- Real-time readingen.rltm.score
- Passage readingen.pred.score
- Spoken choiceen.choc.score
- Semi-open responseen.scne.exam
- Open responseen.prtl.exam
- ASR assessmenten.asr.rec
Mandarin · 8 kernels
- Mandarin charactercn.word.raw · D4
- Mandarin vocabularycn.vocab.raw
- Chinese charactercn.word.raw
- Pinyin assessmentcn.word.score
- Chinese sentencecn.sent.raw
- Chinese passagecn.pred.raw
- Limited-branch recognitioncn.rec.raw
- Chinese AI Talkcn.recscore.raw
/score examples
Score examples
Explore 12 English and 8 Mandarin kernels with illustrative scoring results. Inspect word- or character-level scoring and the documented evidence returned by each core.
en.sent.scoreSentence assessment
Illustrative result using the documented fields available for this kernel.
86Overall
Example response
90–100 Excellent80–89 Good70–79 Developing<70 Retry
Scoring dimensions
Overall86
Accuracy83
Fluency80
Completeness92
Detailed evidence
Pause
2
Speed
186 wpm
Rhythm
65
Word + phoneme
Available
WordScoreFeedback
to74Review pronunciation and phoneme detail before the next attempt.
the64Review pronunciation and phoneme detail before the next attempt.
5 / 12
/structured result
Product-ready evidence — not a black-box percentage.
Overall dimensions arrive with word alignment and phoneme detail. Your app decides thresholds, retries and coaching; Chivox keeps the score deterministic.
Overall 85Accuracy 82Fluency 78Integrity 95
Your app starts the session
SDK call
engine.start({
coreType: 'en.sent.score',
refText: 'An apple a day.'
});Same audio · same reference
Evidence shape
{
overall: 85,
accuracy: 82,
fluency: { overall: 78 },
details: [ /* word + phoneme rows */ ]
}Shared with MCP & function calling
The same engine powers all three integration paths. Switch wrappers later without rebuilding pronunciation logic — only the call contract changes.
/choose-your-integration
SDK, MCP, or function calling?
All three use the same Chivox speech-assessment capability. The right choice depends on who captures the audio, who decides to score, and how much of the loop your application controls.
Current page · SDK & API
| Compare | SDK & API Native scoring inside the product you own. | MCP server Standardized tool layer for compatible clients. | Function calling Typed function inside your own agent stack. |
|---|---|---|---|
| Best-fit scenario | Embed speech assessment inside a learner-facing app or education platform you own. | Expose the same speech tools to MCP-compatible clients and agent environments. | Add speech assessment inside a custom voice agent or an existing LLM workflow. |
| Best-fit teams | EdTech product teams, mobile/web engineers, and platforms that capture audio in-product. | Developer-tool teams, platform teams, and products supporting several MCP clients. | Voice-agent teams, AI application engineers, and product teams with their own orchestration layer. |
| Integration style | Native SDK plus REST/WebSocket — your code starts the session and owns the UX. | Configure one standardized MCP server that publishes discoverable assessment tools. | Define and invoke a typed function contract in your chosen model or agent SDK. |
| Portability | Highest inside your own product surfaces (iOS, Android, web, server); not an agent tool protocol. | Higher across compatible MCP clients; the tool definition stays consistent. | Depends on the provider or framework, but fits deeply into a custom application loop. |
| Application control | Your application controls capture, when to score, thresholds, retries and learner feedback. | The client discovers and calls tools while your server controls assessment behavior and output. | Your application directly controls validation, routing, retries, thresholds, and response handling. |
| Choose it when… | You own the learner journey and need scoring embedded directly in the product UI. | You want the fastest reusable path into MCP ecosystems or need one tool layer for several clients. | You already own the agent loop and need precise control over when and how assessment runs. |
You can mix paths: use the SDK in a learner app, MCP for compatible clients, and function calling inside a proprietary agent — while keeping one shared assessment service underneath.
/workflow
From mic moment to product decision.
Scoring is one step in a product loop. Keep capture UX, thresholds and learner feedback in your code — Chivox supplies the evidence in between.
- 01
Capture in your product
The learner speaks inside your app UI. You decide mic prompts, retakes and when a recording is complete enough to score.
- 02
Send audio + reference
Call the SDK or API with the language, task type, reference text and streamed or submitted audio.
- 03
Receive structured evidence
Overall dimensions arrive with word and phoneme rows — stable fields your product logic can threshold on.
- 04
Drive the next action
Highlight errors, unlock the next drill, ask for a cleaner recording, or store raw evidence for analytics.
Enterprise licensing
SDK & API are licensed separately from MCP plans.
Tell us the platforms, task types and expected traffic. We'll map a pilot, quote licensing and point you at the right docs — usually within one business day.
/contact
Let’s build a better speaking experience together.
Tell us what you’re building and whether you need SDK/API, MCP, or function calling. We’ll reply within one business day with pilot credits, pricing, or a deployment plan.