Add SLNG as Voice AI Infrastructure Provider

providers/slng/models/aura-2.toml

@@ -0,0 +1,27 @@
+# Deepgram Aura 2 TTS (Deepgram-hosted via SLNG)
+# Sources:
+# - https://docs.slng.ai/models — SLNG model catalog (AU, EU, US regions, proxied to Deepgram)
+# - Original per-minute pricing: $0.0085/min converted to per-token at 195 tokens/min
+name = "Aura 2"
+description = "Text-to-speech model supporting 7 languages, proxied to Deepgram's infrastructure"
+family = "aura"
+release_date = "2024-01-01"
+last_updated = "2025-01-01"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 43.59
+output = 43.59
+
+[limit]
+context = 32_768
+output = 16_384
+
+[modalities]
+input = ["text"]
+output = ["audio"]

providers/slng/models/bulbul-v3.toml

@@ -0,0 +1,27 @@
+# Sarvam AI Bulbul v3 TTS via SLNG
+# Sources:
+# - https://docs.slng.ai/models — SLNG model catalog (India region)
+# - Original per-minute pricing: $0.015/min converted to per-token at 195 tokens/min
+name = "Bulbul v3"
+description = "Text-to-speech model supporting 11 Indian and global languages"
+family = "bulbul"
+release_date = "2024-01-01"
+last_updated = "2025-01-01"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 76.92
+output = 76.92
+
+[limit]
+context = 32_768
+output = 16_384
+
+[modalities]
+input = ["text"]
+output = ["audio"]

providers/slng/models/cartesia-sonic-3-5.toml

@@ -0,0 +1,27 @@
+# Cartesia Sonic 3.5 TTS via SLNG
+# Sources:
+# - https://docs.slng.ai/models — SLNG model catalog (EU, India, US regions)
+# - Original per-minute pricing: $0.02/min converted to per-token at 195 tokens/min
+name = "Cartesia Sonic 3.5"
+description = "Advanced text-to-speech model supporting 41 languages with enhanced quality"
+family = "sonic"
+release_date = "2024-01-01"
+last_updated = "2025-01-01"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 102.56
+output = 102.56
+
+[limit]
+context = 32_768
+output = 16_384
+
+[modalities]
+input = ["text"]
+output = ["audio"]

providers/slng/models/cartesia-sonic-3.toml

@@ -0,0 +1,27 @@
+# Cartesia Sonic 3 TTS via SLNG
+# Sources:
+# - https://docs.slng.ai/models — SLNG model catalog (EU, India, US regions)
+# - Original per-minute pricing: $0.02/min converted to per-token at 195 tokens/min
+name = "Cartesia Sonic 3"
+description = "Fast text-to-speech model supporting 39 languages with low latency"
+family = "sonic"
+release_date = "2024-01-01"
+last_updated = "2025-01-01"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 102.56
+output = 102.56
+
+[limit]
+context = 32_768
+output = 16_384
+
+[modalities]
+input = ["text"]
+output = ["audio"]

providers/slng/models/deepgram-aura-2.toml

@@ -0,0 +1,27 @@
+# Deepgram Aura 2 TTS via SLNG
+# Sources:
+# - https://docs.slng.ai/models — SLNG model catalog (US region, self-hosted)
+# - Original per-minute pricing: $0.0085/min converted to per-token at 195 tokens/min
+name = "Deepgram Aura 2"
+description = "Text-to-speech model for natural voice synthesis in English and Spanish"
+family = "aura"
+release_date = "2024-01-01"
+last_updated = "2025-01-01"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 43.59
+output = 43.59
+
+[limit]
+context = 32_768
+output = 16_384
+
+[modalities]
+input = ["text"]
+output = ["audio"]

providers/slng/models/deepgram-nova-3.toml

@@ -0,0 +1,27 @@
+# Deepgram Nova 3 (Multi) STT via SLNG
+# Sources:
+# - https://docs.slng.ai/models — SLNG model catalog (India, AU, US regions, self-hosted)
+# - Original per-minute pricing: $0.0065/min, capped at $3.00 per million tokens
+name = "Deepgram Nova 3 (Multi)"
+description = "Multilingual speech-to-text model supporting English, Spanish, and Hindi"
+family = "nova"
+release_date = "2024-01-01"
+last_updated = "2025-01-01"
+attachment = false
+reasoning = false
+temperature = false
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 3.00
+output = 3.00
+
+[limit]
+context = 0
+output = 0
+
+[modalities]
+input = ["audio"]
+output = ["text"]

providers/slng/models/fish-tts-s2-1-pro-fish.toml

@@ -0,0 +1,27 @@
+# Fish Audio TTS S2.1 Pro (Fish-hosted via SLNG)
+# Sources:
+# - https://docs.slng.ai/models — SLNG model catalog (US region, proxied to Fish Audio)
+# - Fish Audio public pricing: $15.00 per million UTF-8 bytes
+name = "Fish TTS S2.1 Pro"
+description = "Text-to-speech model supporting 23 languages, proxied to Fish Audio"
+family = "fish"
+release_date = "2024-01-01"
+last_updated = "2025-01-01"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 15.00
+output = 15.00
+
+[limit]
+context = 32_768
+output = 16_384
+
+[modalities]
+input = ["text"]
+output = ["audio"]

providers/slng/models/fish-tts-s2-1-pro.toml

@@ -0,0 +1,27 @@
+# Fish Audio TTS S2.1 Pro via SLNG
+# Sources:
+# - https://docs.slng.ai/models — SLNG model catalog (AU, EU regions, self-hosted)
+# - Fish Audio public pricing: $15.00 per million UTF-8 bytes
+name = "Fish TTS S2.1 Pro (SLNG)"
+description = "Multilingual text-to-speech model supporting 50+ languages, hosted on SLNG infrastructure"
+family = "fish"
+release_date = "2024-01-01"
+last_updated = "2025-01-01"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 15.00
+output = 15.00
+
+[limit]
+context = 32_768
+output = 16_384
+
+[modalities]
+input = ["text"]
+output = ["audio"]

providers/slng/models/fish-tts-s2-pro.toml

@@ -0,0 +1,27 @@
+# Fish Audio TTS S2 Pro via SLNG
+# Sources:
+# - https://docs.slng.ai/models — SLNG model catalog (US region, proxied to Fish Audio)
+# - Fish Audio public pricing: $15.00 per million UTF-8 bytes
+name = "Fish TTS S2 Pro"
+description = "Text-to-speech model supporting 23 languages, proxied to Fish Audio"
+family = "fish"
+release_date = "2024-01-01"
+last_updated = "2025-01-01"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 15.00
+output = 15.00
+
+[limit]
+context = 32_768
+output = 16_384
+
+[modalities]
+input = ["text"]
+output = ["audio"]

providers/slng/models/gradium-stt.toml

@@ -0,0 +1,27 @@
+# Gradium STT via SLNG
+# Sources:
+# - https://docs.slng.ai/models — SLNG model catalog (EU region)
+# - Original per-minute pricing: $0.0075/min, capped at $3.00 per million tokens
+name = "Gradium STT"
+description = "Speech-to-text model supporting German, English, French, Italian, and Portuguese"
+family = "gradium"
+release_date = "2024-01-01"
+last_updated = "2025-01-01"
+attachment = false
+reasoning = false
+temperature = false
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 3.00
+output = 3.00
+
+[limit]
+context = 0
+output = 0
+
+[modalities]
+input = ["audio"]
+output = ["text"]

providers/slng/models/gradium-tts.toml

@@ -0,0 +1,27 @@
+# Gradium TTS via SLNG
+# Sources:
+# - https://docs.slng.ai/models — SLNG model catalog (EU region)
+# - Original per-minute pricing: $0.010/min converted to per-token at 195 tokens/min
+name = "Gradium TTS"
+description = "Text-to-speech model supporting German, English, Spanish, French, Japanese, Korean, and Portuguese"
+family = "gradium"
+release_date = "2024-01-01"
+last_updated = "2025-01-01"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 51.28
+output = 51.28
+
+[limit]
+context = 32_768
+output = 16_384
+
+[modalities]
+input = ["text"]
+output = ["audio"]

providers/slng/models/inworld-max-1-5.toml

@@ -0,0 +1,27 @@
+# Inworld Max 1.5 TTS via SLNG
+# Sources:
+# - https://docs.slng.ai/models — SLNG model catalog (EU region, self-hosted)
+# - Original per-minute pricing: $0.010/min converted to per-token at 195 tokens/min
+name = "Inworld Max 1.5"
+description = "Multilingual text-to-speech model supporting 130+ languages"
+family = "inworld"
+release_date = "2024-01-01"
+last_updated = "2025-01-01"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 51.28
+output = 51.28
+
+[limit]
+context = 32_768
+output = 16_384
+
+[modalities]
+input = ["text"]
+output = ["audio"]

providers/slng/models/kugel-2.toml

@@ -0,0 +1,27 @@
+# KugelAudio Kugel 2 TTS via SLNG
+# Sources:
+# - https://docs.slng.ai/models — SLNG model catalog (EU region)
+# - Original per-minute pricing: $0.02/min converted to per-token at 195 tokens/min
+name = "Kugel 2"
+description = "Text-to-speech model supporting 38 languages"
+family = "kugel"
+release_date = "2024-01-01"
+last_updated = "2025-01-01"
+attachment = false
+reasoning = false
+temperature = true
+tool_call = false
+structured_output = false
+open_weights = false
+
+[cost]
+input = 102.56
+output = 102.56
+
+[limit]
+context = 32_768
+output = 16_384
+
+[modalities]
+input = ["text"]
+output = ["audio"]