Skip to main content
Glama
elibarnett

kie-mcp

by elibarnett
OverviewSchemaSourceRelated ServersDiscussionsAdmin
JavaScript
Hybrid

generate_gemini_tts

Convert text into AI speech with Google Gemini TTS, supporting single-speaker narration or up to two-speaker dialogue with style-directed delivery and named voices.

Instructions

NEW — Google Gemini native TTS via kie.ai: style-directed speech from natural-language direction, 30 named voices, up to 2 speakers, inline tone tags like [whispers]/[laughs] (flash model). ~4.2 credits per MINUTE of audio — cheaper than all ElevenLabs tiers. Simple mode: pass text (+ optional voice_name). Dialogue mode: pass speakers + dialogue_turns. model=flash is most expressive (keep expected audio <60s — quality degrades on long takes); model=pro is more stable for multi-minute narration. Downloads to kie/assets/raw/.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
textNoSimple mode: the text to speak (single speaker). Inline tone tags like [whispers] work on flash. Ignored if dialogue_turns is set.
waitNoSet false to submit and return immediately with the task_id (async mode) — then poll with check_task and fetch with download_result.
modelNoflash = Gemini 3.1 Flash TTS (most expressive, 200+ inline tags, best <60s; ~4.2 cr/min). flash-3.8 = Gemini 3.8 Flash TTS (NEW Sept 2026, same inputs, ~1.9 cr/min, ~40% cheaper per clip in live tests). flash-lite = Gemini 3.8 Flash Lite TTS (NEW, cheapest at ~1.3 cr/min; draft/bulk narration). pro = Gemini 2.5 Pro TTS (more stable long-form, ~4.2 cr/min). 3.8 prices are kie limited-time pricing until 2026-12-31.flash
sceneNoScene description, e.g. "A quiet, warm room with a fireplace crackling softly."
filenameNoOutput filename (default gemini-tts-<ts>.wav)
speakersNoDialogue mode: 1-2 speakers as [{speaker_id: "Speaker 1", voice_name, audio_profile?, accent?, style?, pace?}]. accent: Neutral|American (Gen)|American (Valley)|American (South)|British (RP)|British (Brixton)|Transatlantic|Australian. style: Vocal Smile|Newscaster|Whisper|Empathetic|Promo/Hype|Deadpan. pace: Natural|Rapid Fire|The Drift|Staccato.
voice_nameNoSimple mode voice (default Zephyr)
temperatureNoSampling temperature (default 1)
download_dirNoAbsolute directory to save into (created if missing). Defaults to the server's kie/assets/raw/.
dialogue_turnsNoDialogue mode: [{speaker_id, text}] in order; text ≤10000 chars, may contain tone tags
sample_contextNoOverall tone/direction, e.g. "Audiobook style narration. Tone is gentle and inviting."
max_wait_secondsNo

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed2 schema fields changedv5.3.0
    • changedInput schema / properties / model / description
      Previous value: -"flash = Gemini 3.1 Flash TTS (most expressive, 200+ inline tags, best <60s); pro = Gemini 2.5 Pro TTS (more stable long-form). Same price."New value: +"flash = Gemini 3.1 Flash TTS (most expressive, 200+ inline tags, best <60s; ~4.2 cr/min). flash-3.8 = Gemini 3.8 Flash TTS (NEW Sept 2026, same inputs, ~1.9 cr/min, ~40% cheaper per clip in live tests). flash-lite = Gemini 3.8 Flash Lite TTS (NEW, cheapest at ~1.3 cr/min; draft/bulk narration). pro = Gemini 2.5 Pro TTS (more stable long-form, ~4.2 cr/min). 3.8 prices are kie limited-time pricing until 2026-12-31."
    • changedInput schema / properties / model / enum
      Previous value: -[
-  "flash",
-  "pro"
-]New value: +[
+  "flash",
+  "flash-3.8",
+  "flash-lite",
+  "pro"
+]
  2. First observedv4.7.0

TDQS

A3.9/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries the full burden and does well: it discloses cost (~4.2 credits/min, cheaper than ElevenLabs), the quality-degradation risk on long flash takes, the output destination (kie/assets/raw/), and limited-time pricing expiry. It omits nothing critical, though it says nothing about auth/permissions or concurrency limits.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with the identity and the two operating modes, and the model tradeoffs are tightly packed. Some promotional phrasing ('cheaper than all ElevenLabs tiers', repeated pricing) consumes space without helping invocation, keeping it from a 5.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 12-parameter tool with no annotations and no output schema, the description covers mode routing, model tradeoffs, cost, and save location well. It leaves speaker_id/pacing semantics and the async polling flow to the schema, which is acceptable given 92% coverage.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is already 92%, so the schema documents nearly every parameter (model enums, wait, speakers, temperature). The description largely restates schema content (2-speaker cap, tone tags, enum meaning) rather than adding new semantic detail, which is the baseline-3 case.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb+resource (native Gemini TTS via kie.ai) and enumerates the distinguishing capabilities: style-directed speech, 30 named voices, up to 2 speakers, inline tone tags. It never names sibling tools like generate_tts or generate_dialogue, so the agent must infer the boundary rather than being routed explicitly.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives clear mode selection rules (simple mode = text+voice_name; dialogue mode = speakers+dialogue_turns) and model selection guidance keyed to clip length (flash for ~60s expressive takes, pro for stable multi-minute narration). However, it offers no when-not-to-use this tool versus the sibling TTS/dialogue tools.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Deploy Server

Other Tools