MCP Fish Server

fish_generate_audio

Convert text into natural-sounding speech audio with AI voice synthesis. Choose a voice reference or voice ID to generate spoken narration for any text.

Instructions

Generate speech audio from text using Fish TTS.

Use this when:
- You want to convert text to speech
- You need AI-generated audio narration

Returns:
    JSON response containing task_id and audio data when complete.

Example:
    fish_generate_audio(
        text="Hello, welcome to our service!",
        reference_id="d7900c21663f485ab63ebdb7e5905036"
    )

Input Schema

TableJSON Schema

Name	Required	Description	Default
`text`	No	The text to synthesize. Required.
`async`	No	Whether to process the request asynchronously.
`model`	No	The TTS model to use. Supported values: 's1', 's2-pro', 's2.1-pro'.	s2-pro
`top_p`	No	Top-p nucleus sampling parameter.
`format`	No	Output audio format. Defaults to 'mp3'.	mp3
`prompt`	No	Deprecated alias for `text`.
`latency`	No	Latency mode. Supported values: 'normal', 'balanced'.
`prosody`	No	Prosody overrides forwarded to the upstream.
`voice_id`	No	Deprecated alias for `reference_id`.
`normalize`	No	Whether the upstream should apply text normalization.
`references`	No	One-shot voice clone reference: one object with a public HTTPS audio URL and the exact transcript. Cannot be combined with reference_id.
`mp3_bitrate`	No	MP3 bit rate when format='mp3'. Supported values: 64, 128, 192.
`sample_rate`	No	Sampling rate of the output audio (e.g. 16000, 22050, 44100).
`temperature`	No	Sampling temperature (0.0-1.0).
`callback_url`	No	Optional callback URL to receive the result asynchronously. If provided, the API returns immediately with a task_id.
`chunk_length`	No	Chunk length passed through to the upstream synthesiser.
`reference_id`	No	Voice model id, or a list of voice model ids. Default example voice model ID: d7900c21663f485ab63ebdb7e5905036.
`max_new_tokens`	No	Maximum number of new tokens to generate.
`reference_text`	No	Exact transcript for reference_audio_url.
`min_chunk_length`	No	Minimum chunk length.
`repetition_penalty`	No	Repetition penalty applied during generation.
`reference_audio_url`	No	Convenience HTTPS audio URL for one-shot voice cloning.

Output Schema

TableJSON Schema

Name	Required	Description	Default
`result`	Yes

Schema Changelog

Changes observed during successful MCP inspections. Dates show when Glama detected each change.

Changed7 schema fields changedv0.1.3Sep 6, 2026

addedInput schema / $defs

Added value: +{
+  "FishReference": {
+    "description": "One-shot voice reference hosted at a public HTTPS URL.",
+    "properties": {
+      "audio": {
+        "format": "uri",
+        "maxLength": 2083,
+        "minLength": 1,
+        "title": "Audio",
+        "type": "string"
+      },
+      "text": {
+        "maxLength": 10000,
+        "minLength": 1,
+        "title": "Text",
+        "type": "string"
+      }
+    },
+    "required": [
+      "audio",
+      "text"
+    ],
+    "title": "FishReference",
+    "type": "object"
+  }
+}

addedInput schema / properties / reference_audio_url

Added value: +{
+  "anyOf": [
+    {
+      "format": "uri",
+      "maxLength": 2083,
+      "minLength": 1,
+      "type": "string"
+    },
+    {
+      "type": "null"
+    }
+  ],
+  "default": null,
+  "description": "Convenience HTTPS audio URL for one-shot voice cloning.",
+  "title": "Reference Audio Url"
+}

changedInput schema / properties / reference_id / anyOf

Previous value: -[
-  {
-    "type": "string"
-  },
-  {
-    "type": "null"
-  }
-]New value: +[
+  {
+    "type": "string"
+  },
+  {
+    "items": {
+      "type": "string"
+    },
+    "type": "array"
+  },
+  {
+    "type": "null"
+  }
+]

changedInput schema / properties / reference_id / description

Previous value: -"Voice model id (single speaker). Default example voice model ID: d7900c21663f485ab63ebdb7e5905036."New value: +"Voice model id, or a list of voice model ids. Default example voice model ID: d7900c21663f485ab63ebdb7e5905036."

addedInput schema / properties / reference_text

Added value: +{
+  "anyOf": [
+    {
+      "maxLength": 10000,
+      "minLength": 1,
+      "type": "string"
+    },
+    {
+      "type": "null"
+    }
+  ],
+  "default": null,
+  "description": "Exact transcript for reference_audio_url.",
+  "title": "Reference Text"
+}

changedInput schema / properties / references / anyOf

Previous value: -[
-  {
-    "items": {
-      "additionalProperties": true,
-      "type": "object"
-    },
-    "type": "array"
-  },
-  {
-    "type": "null"
-  }
-]New value: +[
+  {
+    "items": {
+      "$ref": "#/$defs/FishReference"
+    },
+    "maxItems": 1,
+    "minItems": 1,
+    "type": "array"
+  },
+  {
+    "type": "null"
+  }
+]

changedInput schema / properties / references / description

Previous value: -"Inline reference samples forwarded to the upstream."New value: +"One-shot voice clone reference: one object with a public HTTPS audio URL and the exact transcript. Cannot be combined with reference_id."

Changed2 schema fields changedv0.1.1Aug 21, 2026

addedInput schema / properties / async

Added value: +{
+  "anyOf": [
+    {
+      "type": "boolean"
+    },
+    {
+      "type": "null"
+    }
+  ],
+  "default": null,
+  "description": "Whether to process the request asynchronously.",
+  "title": "Async"
+}

changedInput schema / properties / mp3_bitrate / anyOf

Previous value: -[
-  {
-    "type": "integer"
-  },
-  {
-    "type": "null"
-  }
-]New value: +[
+  {
+    "enum": [
+      64,
+      128,
+      192
+    ],
+    "type": "integer"
+  },
+  {
+    "type": "null"
+  }
+]

First observedv0.1.0Aug 1, 2026

TDQS

B3.4/5.0

Behavior2/5

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

No annotations are provided, so the description carries the full burden of behavioral disclosure. It mentions the JSON return shape but omits critical behavior such as async/callback_url returning a task_id before audio is ready and the need to poll fish_get_task for results. This is a significant transparency gap for a complex synthesis API.

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

Conciseness3/5

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

The first sentence is an efficient purpose statement and the example is useful. However, the 'Use this when' bullets largely restate the first sentence ('convert text to speech' is nearly identical to 'generate speech audio from text'), adding redundancy without new information.

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

Completeness2/5

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

This is a complex tool with 22 optional parameters, asynchronous modes, and voice-cloning options, yet the description only covers the basic synchronous path. It omits operational guidance for async/callback_url, references/reference_audio_url, and retrieving results via fish_get_task, so an agent could easily misuse the tool for non-default use cases.

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 description coverage is 100%, so the baseline is 3. The example adds practical semantics for text and reference_id, but the description does not explain parameter relationships like references vs reference_id or async vs callback_url, which the schema also leaves implicit.

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

Purpose5/5

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

States a specific action ('Generate speech audio from text') and the service ('Fish TTS'), which clearly distinguishes it from the sibling get/list tools. The 'Use this when' bullets reinforce the purpose and leave no doubt about what the tool is for.

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?

Provides explicit 'Use this when' conditions for text-to-speech and narration, giving clear invocation context. It does not name alternatives or exclusions, but no sibling tool generates audio, so the guidance is sufficient.

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

Install Server

Other Tools

Latest Blog Posts

Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)
By Om-Shree-0709 on July 25, 2026.
mcp
Agent Identity
OAuth 2.1
Your AI Chatbot Just Exposed Your CEO's Salary to an Intern
By Om-Shree-0709 on July 2, 2026.
Agent Identity
MCP Security
OAuth Delegation
Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)
By Om-Shree-0709 on June 30, 2026.
Agentic Ai
Prompt Injection
WebAssembly

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/AceDataCloud/FishMCP'

If you have feedback or need assistance with the MCP directory API, please join our Discord server