CtrlK
BlogDocsLog inGet started
Tessl Logo

deepgram-java-speech-to-text

Use when writing or reviewing Java code in this repo that calls Deepgram Speech-to-Text v1 (`/v1/listen`) for prerecorded or live transcription. Covers `client.listen().v1().media().transcribeUrl` / `transcribeFile` (REST) and `client.listen().v1().v1WebSocket()` (WebSocket). Use `deepgram-java-audio-intelligence` for analytics overlays, `deepgram-java-conversational-stt` for Flux `/v2/listen`, and `deepgram-java-voice-agent` for full-duplex assistants. Triggers include "transcribe", "speech to text", "STT", "listen v1", "nova-3", "live transcription", and "websocket transcription".

74

Quality

91%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Low

Low-risk findings worth noting

SKILL.md
Quality
Evals
Security

Quality

Content

82%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

An executable, well-organized reference skill with copy-paste-ready Java examples for REST, WebSocket, and async paths plus a strong Gotchas section. The main lift would be moving the dense API-surface catalog into a separate reference file to improve progressive disclosure and token efficiency.

Suggestions

Move the 'Key parameters / API surface' catalog (lines 162-168) into a separate reference file and link to it, so the inline body stays lean and progressive disclosure reaches the top anchor.

Trim or relocate the 'Central product skills' promo (lines 200-208) — it reads as marketing bulk rather than skill-internal guidance and competes with the context budget.

For the WebSocket quick start, add an explicit validation checkpoint (e.g., verify `connect(...).get(...)` succeeded before sending media) so the live-streaming sequence has a clear checkpoint.

DimensionReasoningScore

Conciseness

Dense, assumes Claude's competence (no explaining what Java/REST/libraries are), but the inline 'Key parameters / API surface' list and the 'Central product skills' promo add bulk that could be externalized; a 5 would have every token earning its place with no reference-style padding.

4 / 5

Actionability

Multiple copy-paste-ready executable Java blocks (REST URL with full visitor handling, REST file bytes, WebSocket, async) with real imports and class paths verified against the repo; covers common cases fully, so above the 4 'minor gaps' anchor.

5 / 5

Workflow Clarity

Quick starts are clear sequences and the Gotchas section encodes ordering constraints (register handlers before connect, wait on connect().get()); no destructive/batch ops trigger the validation cap, but as a reference skill it lacks explicit multi-step checkpoints that would justify a 5.

4 / 5

Progressive Disclosure

No bundle files exist, but in-repo source paths and external canonical OpenAPI/AsyncAPI URLs are clearly signaled and one-level deep with a layered 1-5 reference list; the inline API-surface catalog (lines 162-168) is content that could live in a separate reference file, keeping it just below the 5 'well-signaled one-level' anchor.

4 / 5

Total

17

/

20

Passed

Description

100%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A highly specific, well-triggered description with explicit 'what' and 'when' guidance and clear disambiguation from sibling Deepgram skills. It avoids verbosity while remaining comprehensive.

DimensionReasoningScore

Specificity

Names the domain (Java code calling Deepgram STT v1) plus concrete methods — `transcribeUrl`/`transcribeFile` (REST) and `v1WebSocket()` (WebSocket) — distinguishing prerecorded vs live; comprehensive concrete actions, not a 4 because coverage of capabilities is complete rather than having minor gaps.

5 / 5

Completeness

Clearly answers 'what' (Java SDK calls to `/v1/listen` over REST and WebSocket) and 'when' via an explicit 'Use when writing or reviewing Java code...' clause with concrete trigger phrases; not below 5 because both halves are explicit and concrete.

5 / 5

Trigger Term Quality

Explicit trigger list 'transcribe', 'speech to text', 'STT', 'listen v1', 'nova-3', 'live transcription', 'websocket transcription' covers natural user phrasings and synonyms; not a 4 because both common terms and model/transport synonyms are present.

5 / 5

Distinctiveness Conflict Risk

Explicitly redirects to `deepgram-java-audio-intelligence`, `deepgram-java-conversational-stt`, and `deepgram-java-voice-agent` with distinct boundary criteria, giving a clear niche with minimal conflict risk.

5 / 5

Total

20

/

20

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository
deepgram/deepgram-java-sdk
Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.