togethercomputer/together-typescript v0.40.0

v0.40.0

Repository: togethercomputer/together-typescript

Tag: v0.40.0

Published: 2026-05-11T14:54:43Z

Prerelease: no

Release notes:

0.40.0 (2026-05-11)

Full Changelog: v0.39.0...v0.40.0

Features

• add the pronunciation dict (4f60c77)
• api: accept string URLs for audio_inputs and source_video in videos (9313762)
• api: add cached_input field to model pricing (497b932)
• api: add cuda/nvidia driver fields, auto-scaling, OIDC, scheduled capacity to clusters (881242f)
• api: add h100-40gb-mig and b200-192gb gpu_type options to jig resource (dcadaf2)
• api: add max_seq_length parameter to fine-tuning (2dd78f9)
• api: add max_tokens and temperature params to evals judge model config (5471896)
• api: add num_workers parameter to evals model/judge requests (c61a7c1)
• api: remove task field from audio transcriptions/translations responses (8125896)
• api: Update server url to .ai (b642c6c)
• ENG-87042: clarify TTS language parameter supports lowercase locales (914349f)
• ENG-87042: document language on TTS WebSocket and simplify locale note (bb4b3f4)
• MOSH-2181: Add default note on max-seq-length (9fce848)
• support setting headers via env (e649d67)

Bug Fixes

• client: update videos base URL to api.together.ai (fcf6e84)
• types: remove enum values from FilePurpose (0f538e0)
• types: rename cuda_driver_version to cuda_version in clusters (e83d0c4)
• types: update driver_versions structure, require supported_instance_types in cluster regions (d1caf1c)

Chores

• format: run eslint and prettier separately (0fcf0fe)
• formatter: run prettier and eslint separately (fabef51)
• internal: codegen related update (6271f5e)
• internal: codegen related update (f9b079e)
• internal: codegen related update (4ba1386)
• internal: more robust bootstrap script (5c001e2)
• redact api-key headers in debug logs (dff31c6)
• tests: bump steady to v0.22.1 (3d2bf61)

Documentation

• api: document voice mixing support for speech voice parameter (12e01c6)
• api: expand billing_type documentation in cluster creation (f74418e)
• api: update prompt parameter description in audio transcriptions/translations (00c887c)
• api: update response_encoding and sample_rate descriptions in speech (44d908d)
• api: update supported file formats in audio transcriptions/translations (824c080)
• improve examples (bdf968a)