• 0.1.0latestno API changesDocs2026-07-17 12:04
• update openai api, in accordance to the work done in the atoma proxy (#408)38841Docs00da4952025-02-12 22:34

  Change detail

  87 changes across 4 operations and 7 models

  AllBreakingWarningInfo

  • POST/v1/chat/completions
    • breaking

      removed StopCondition subschema #1 from the stop request property oneOf list

    • breaking

      removed subschema #1 subschema #2 subschema #3 subschema #4 from the messages/items request property oneOf list

    • breaking

      added 'propertyNames' constraint to the request property tools/items/function/parameters

    • breaking

      the logit_bias/propertyNames request property type changed from string to integer, and format from no format to int32

    • breaking

      the n request property format changed from no format to int32

    • breaking

      the stop request property type changed from no type to array null

    • breaking

      the tools/items/function/parameters request property type changed from no type to object null

  View all changes (87)
• feat: deprecate the use of max_tokens to max_completion_tokens, following openai api spec (#378)4Docs3c54fee2025-01-29 21:10

  Change detail

  4 changes across 2 operations

  AllBreakingWarningInfo

  • POST/v1/chat/completions
    • info

      added the new optional request property max_completion_tokens

    • info

      request property max_tokens deprecated

  • POST/v1/confidential/chat/completions
    • info

      added the new optional request property max_completion_tokens

    • info

      request property max_tokens deprecated

• feat: add support for multiple chat completions services (#355)no API changesDocs985c7f42025-01-27 11:14
• feat: refactor confidential compute endpoints (#319)24Docs5598e582025-01-03 10:33

  Change detail

  6 changes across 4 operations

  AllBreakingWarningInfo

  • POST/v1/chat/completions
    • breaking

      the response property usage/completion_tokens became nullable for the status 200

    • breaking

      the response property usage/completion_tokens became optional for the status 200

    • info

      added the optional property usage/completion_tokens_details to the response with the 200 status

  • POST/v1/confidential/chat/completions
    • info

      endpoint added

  • POST/v1/confidential/embeddings
    • info

      endpoint added

  • POST/v1/confidential/images/generations
    • info

      endpoint added

• first commit (#315)no API changesDocs2a21a6b2024-12-30 10:15
• feat: add prometheus docker service for node performance metrics analytics (#287)no API changesDocs4dda3aa2024-12-16 10:57
• fix: resolve bugs with confidential compute (#268)1Docs5c5fffb2024-12-06 07:22

  Change detail

  1 change across 1 operation

  AllBreakingWarningInfo

  • GET/metrics
    • info

      endpoint added

• fix: proper shutdown/failure (#239)11Docs8996a472024-11-15 15:04

  Change detail

  2 changes across 1 operation

  AllBreakingWarningInfo

  • POST/v1/chat/completions
    • warning

      removed the request property max_completion_tokens

    • info

      added the new optional request property max_tokens

• feat: inference service endpoints (#217)131541Docs22aa29b2024-11-12 11:48

Revisions go back to the API's first capture.