elevenlabs 0.0.5 → 0.0.7

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 59c61b4f80cd6efaaa39b93d6ab8ebfdc3a3edce59d2012a984ce8b9f032352d
-  data.tar.gz: 06c3253d6d21cd59fa5942620de9dee173a3ac926c484dcd17ca378793f6920e
+  metadata.gz: 2daafae7b6dbf3724b93ce2022b2fe6ac3703bfbcac12326b75e1a37cd188a39
+  data.tar.gz: ba2227a765efc7538e4aadbe0fcb0917a55c1ba70540a2660b4c75b2545f85da
 SHA512:
-  metadata.gz: 5f1cd0d00b602fe88356591d3292dfc6154d722eab9904dea0ec0cd99a5340651be30f44e0470783c1504cabeeed4a0e4dcfab6732c742a650c3c4161129c775
-  data.tar.gz: b9e3f4821d6e355f811f0d783b36cec64dd103a83713eb13a3ff3ae8c886009294371df5de825ff9287be969b92489b19c90ce426584a81ddb34d8f6e5e97182
+  metadata.gz: 07d40969dd5fdf8926c2f09c21359df4b5060b1f212797de03ef16c4fcf0dc2b6495c476a9a0741247dccf95dd30fde8d6da7404370b6acc4b61a0ea0ce8f7cd
+  data.tar.gz: 927e01fdc01e4f985466b62e2725f676117f757d3f809d8b4da7ea420e54c0e0a57ab2279cf031dd466451ce8c7a96dd76219da4a7e3d02ada842682a20246f8

data/README.md

@@ -14,6 +14,7 @@ This gem provides an easy-to-use interface for:
 - **Converting text to speech** and retrieving the generated audio
 - **Designing a voice** based on a text description
 - **Streaming text-to-speech audio**
+- **Music Generation**
 
 All requests are handled via [Faraday](https://github.com/lostisland/faraday).
 
@@ -196,16 +197,20 @@ end
 ```ruby
 client.list_voices
 # => { "voices" => [...] }
-```
 
-2. **Get Voice Details**
+2. List Models
+
+client.list_models
+# => [...]
+
+3. **Get Voice Details**
 
 ```ruby
 client.get_voice("VOICE_ID")
 # => { "voice_id" => "...", "name" => "...", ... }
 ```
 
-3. **Create a Custom Voice**
+4. **Create a Custom Voice**
 
 ```ruby
 sample_files = [File.open("sample1.mp3", "rb")]
@@ -213,7 +218,7 @@ client.create_voice("Custom Voice", sample_files, description: "My custom AI voi
 # => JSON response with new voice details
 ```
 
-4. **Check if a Voice is Banned**
+5. **Check if a Voice is Banned**
 
 ```ruby
 sample_files = [File.open("trump.mp3", "rb")]
@@ -224,28 +229,28 @@ client.banned?(trump)
 # => true
 ```
 
-5. **Edit a Voice**
+6. **Edit a Voice**
 
 ```ruby
 client.edit_voice("VOICE_ID", name: "Updated Voice Name")
 # => JSON response with updated details
 ```
 
-6. **Delete a Voice**
+7. **Delete a Voice**
 
 ```ruby
 client.delete_voice("VOICE_ID")
 # => JSON response acknowledging deletion
 ```
 
-7. **Convert Text to Speech**
+8. **Convert Text to Speech**
 
 ```ruby
 audio_data = client.text_to_speech("VOICE_ID", "Hello world!")
 File.open("output.mp3", "wb") { |f| f.write(audio_data) }
 ```
 
-8. **Stream Text to Speech**
+9. **Stream Text to Speech**
 
 Stream from terminal:
 
@@ -264,22 +269,75 @@ IO.popen("play -t mp3 -", "wb") do |audio_pipe| # Notice "wb" (write binary)
 end
 ```
 
-9. **Design a Voice**
+10. **Create a Voice from a Design**
 
-Generate voice previews based on a text description:
+Once you’ve generated a voice design using client.design_voice, you can turn it into a permanent voice in your account by passing its generated_voice_id to client.create_from_generated_voice.
 
+# Step 1: Design a voice (returns previews + generated_voice_id)
 ```ruby
-response = client.design_voice(
-  "A deep, resonant male voice with a British accent, suitable for storytelling",
-  output_format: "mp3_44100_192",
+design_response = client.design_voice(
+  "A warm, friendly female voice with a slight Australian accent",
   model_id: "eleven_multilingual_ttv_v2",
-  text: "In a land far away, where the mountains meet the sky, a great adventure began. Brave heroes embarked on a quest to find the lost artifact, facing challenges and forging bonds that would last a lifetime. Their journey took them through enchanted forests, across raging rivers, and into the heart of ancient ruins.",
+  text: "Welcome to our podcast, where every story is an adventure, taking you on a journey through fascinating worlds, inspiring voices, and unforgettable moments.",
   auto_generate_text: false
 )
-# Save the first preview to an MP3 file
-require "base64"
-audio_data = Base64.decode64(response["previews"][0]["audio_base_64"])
-File.open("voice_preview.mp3", "wb") { |f| f.write(audio_data) }
+
+generated_voice_id = design_response["previews"].first["generated_voice_id"] #three previews are given, but for this example we will use the first to create a voice here
+
+# Step 2: Create the permanent voice
+create_response = client.create_from_generated_voice(
+  "Friendly Aussie",
+  "A warm, friendly Australian-accented voice for podcasts",
+   generated_voice_id,
+)
+
+voice_id = create_response["voice_id"] # This is the ID you can use for TTS
+
+# Step 3: Use the new voice for TTS
+audio_data = client.text_to_speech(voice_id, "This is my new permanent designed voice.")
+File.open("friendly_aussie.mp3", "wb") { |f| f.write(audio_data) }
+```
+Important notes:
+
+Always store the returned voice_id from create_voice_from_design. This is the permanent identifier for TTS.
+
+Designed voices cannot be used for TTS until they are created in your account.
+
+If the voice is not immediately available for TTS, wait a few seconds or check its status via client.get_voice(voice_id) until it’s "active".
+
+11. Create a multi-speaker dialogue
+```ruby
+inputs = [{text: "It smells like updog in here", voice_id: "TX3LPaxmHKxFdv7VOQHJ"}, {text: "What's updog?", voice_id: "RILOU7YmBhvwJGDGjNmP"}, {text: "Not much, you?", voice_id: "TX3LPaxmHKxFdv7VOQHJ"}]
+
+audio_data = client.text_to_dialogue(inputs)
+File.open("what's updog.mp3", "wb") { |f| f.write(audio_data) }
+```
+
+12. **Generate Music from prompt**
+```ruby
+audio = client.compose_music(prompt: "Lo-fi hip hop beat", music_length_ms: 30000)
+File.binwrite("lofi.mp3", audio)
+```
+
+12. **Stream Music Generated from prompt**
+```ruby
+File.open("epic_stream.mp3", "wb") do |f|
+  client.compose_music_stream(prompt: "Epic orchestral build", music_length_ms: 60000) do |chunk|
+    f.write(chunk)
+  end
+end
+```
+
+13. **Generate Music with Detailed Metadata (metadata + audio) from prompt**
+```ruby
+result = client.compose_music_detailed(prompt: "Jazz piano trio", music_length_ms: 20000)
+puts result # raw multipart data (needs parsing)
+```
+
+14. **Create a music composition plan from prompt**
+```ruby
+plan = client.create_music_plan(prompt: "Upbeat pop song with verse and chorus", music_length_ms: 60000)
+puts plan[:sections]
 ```
 
 ---
@@ -338,7 +396,7 @@ gem build elevenlabs.gemspec
 Install the gem locally:
 
 ```bash
-gem install ./elevenlabs-0.0.5.gem
+gem install ./elevenlabs-0.0.7.gem
 ```
 
 ---

data/lib/elevenlabs/client.rb

@@ -88,6 +88,47 @@ module Elevenlabs
       handle_error(e)
     end
 
+    #####################################################
+    #              Text-to-Dialogue                     #
+    #    (POST /v1/text-to-dialogue)                    #
+    #####################################################
+
+    # Converts a list of text and voice ID pairs into speech (dialogue) and returns audio.
+    # Documentation: https://elevenlabs.io/docs/api-reference/text-to-dialogue/convert
+    #
+    # @param [Array[Objects]] inputs - A list of dialogue inputs, each containing text and a voice ID which will be converted into speech
+    #   :text => String
+    #   :voice_id => String
+    # @param [String] model_id - optional Identifier of the model to be used
+    # @param [Hash] settings - optinal Settings controlling the dialogue generation
+    #   :stability => double - 0.0 = Creative, 0.5 = Natural, 1.0 = Robust
+    #   :use_speaker_boost => boolean
+    # @param [Integer] seed - optional Best effort to sample deterministically.
+    #
+    # @return [String] The binary audio data (usually an MP3).
+    def text_to_dialogue(inputs, model_id = nil, settings = {}, seed = nil)
+      endpoint = "/v1/text-to-dialogue"
+      request_body = {}.tap do |r|
+        r[:inputs] = inputs
+        r[:model_id] = model_id if model_id
+        r[:settings] = settings unless settings.empty?
+        r[:seed] = seed if seed
+      end
+
+      headers = default_headers
+      headers["Accept"] = "audio/mpeg"
+
+      response = @connection.post(endpoint) do |req|
+        req.headers = headers
+        req.body = request_body.to_json
+      end
+
+      # Returns raw binary data (often MP3)
+      response.body
+    rescue Faraday::ClientError => e
+      handle_error(e)
+    end
+
     #####################################################
     #                  Design a Voice                   #
     #      (POST /v1/text-to-voice/design)              #
@@ -194,6 +235,25 @@ module Elevenlabs
       handle_error(e)
     end
 
+    #####################################################
+    #                     GET models #
+    #                  (GET /v1/models)                 #
+    #####################################################
+
+    # Gets a list of available models
+    # Documentation: https://elevenlabs.io/docs/api-reference/models/list
+    #
+    # @return [Hash] The JSON response containing an array of models
+    def list_models
+      endpoint = "/v1/models"
+      response = @connection.get(endpoint) do |req|
+        req.headers = default_headers
+      end
+      JSON.parse(response.body)
+    rescue Faraday::ClientError => e
+      handle_error(e)
+    end
+
     #####################################################
     #                 GET a Single Voice                #
     #               (GET /v1/voices/{voice_id})         #
@@ -345,6 +405,110 @@ module Elevenlabs
       voice_id.in?(active_voices)
     end
 
+    #####################################################
+    #                     Music API                     #
+    #####################################################
+
+    # 1. Compose music (basic)
+    # POST /v1/music
+    def compose_music(options = {})
+      endpoint = "/v1/music"
+      request_body = {
+        prompt: options[:prompt],
+        composition_plan: options[:composition_plan],
+        music_length_ms: options[:music_length_ms],
+        model_id: options[:model_id] || "music_v1"
+      }.compact
+
+      headers = default_headers.merge("Accept" => "audio/mpeg")
+      query = {}
+      query[:output_format] = options[:output_format] if options[:output_format]
+
+      response = @connection.post("#{endpoint}?#{URI.encode_www_form(query)}") do |req|
+        req.headers = headers
+        req.body = request_body.to_json
+      end
+
+      response.body # raw binary audio
+    rescue Faraday::ClientError => e
+      handle_error(e)
+    end
+
+    # 2. Stream music
+    # POST /v1/music/stream
+    def compose_music_stream(options = {}, &block)
+      endpoint = "/v1/music/stream"
+      request_body = {
+        prompt: options[:prompt],
+        composition_plan: options[:composition_plan],
+        music_length_ms: options[:music_length_ms],
+        model_id: options[:model_id] || "music_v1"
+      }.compact
+
+      headers = default_headers.merge("Accept" => "audio/mpeg")
+      query = {}
+      query[:output_format] = options[:output_format] if options[:output_format]
+
+      @connection.post("#{endpoint}?#{URI.encode_www_form(query)}") do |req|
+        req.options.on_data = Proc.new do |chunk, _|
+          block.call(chunk) if block
+        end
+        req.headers = headers
+        req.body = request_body.to_json
+      end
+
+      nil # audio streamed via block
+    rescue Faraday::ClientError => e
+      handle_error(e)
+    end
+
+    # 3. Compose detailed music (metadata + audio)
+    # POST /v1/music/detailed
+    def compose_music_detailed(options = {})
+      endpoint = "/v1/music/detailed"
+      request_body = {
+        prompt: options[:prompt],
+        composition_plan: options[:composition_plan],
+        music_length_ms: options[:music_length_ms],
+        model_id: options[:model_id] || "music_v1"
+      }.compact
+
+      headers = default_headers
+      query = {}
+      query[:output_format] = options[:output_format] if options[:output_format]
+
+      response = @connection.post("#{endpoint}?#{URI.encode_www_form(query)}") do |req|
+        req.headers = headers
+        req.body = request_body.to_json
+      end
+
+      response.body # multipart/mixed with JSON + binary audio
+    rescue Faraday::ClientError => e
+      handle_error(e)
+    end
+
+
+    # 4. Create a composition plan
+    # POST /v1/music/plan
+    def create_music_plan(options = {})
+      endpoint = "/v1/music/plan"
+      request_body = {
+        prompt: options[:prompt],
+        music_length_ms: options[:music_length_ms],
+        source_composition_plan: options[:source_composition_plan],
+        model_id: options[:model_id] || "music_v1"
+      }.compact
+
+      response = @connection.post(endpoint) do |req|
+        req.headers = default_headers
+        req.body = request_body.to_json
+      end
+
+      JSON.parse(response.body, symbolize_names: true)
+    rescue Faraday::ClientError => e
+      handle_error(e)
+    end
+
     private
 
     # Common headers needed by Elevenlabs

data/lib/elevenlabs.rb

@@ -5,7 +5,7 @@ require_relative "elevenlabs/client"
 require_relative "elevenlabs/errors"
 
 module Elevenlabs
-  VERSION = "0.0.5"
+  VERSION = "0.0.7"
 
   # Optional global configuration
   class << self

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: elevenlabs
 version: !ruby/object:Gem::Version
-  version: 0.0.5
+  version: 0.0.7
 platform: ruby
 authors:
 - hackliteracy
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2025-08-09 00:00:00.000000000 Z
+date: 2025-08-25 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: faraday
@@ -39,7 +39,8 @@ dependencies:
       - !ruby/object:Gem::Version
         version: '1.1'
 description: This gem provides a convenient Ruby interface to the ElevenLabs TTS,
-  Voice Cloning, Voice Design and Streaming endpoints.
+  Voice Cloning, Voice Design, Voice dialogues, TTS Streaming, Music Generation and
+  Streaming endpoints.
 email:
 - hackliteracy@gmail.com
 executables: []