Skip to content

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[0.95.0]

  • Module changes:
    • Anthropic
      • Invoke-AnthropicModel
        • Added support for two new models: anthropic.claude-opus-4-20250514-v1:0, anthropic.claude-sonnet-4-20250514-v1:0

[0.90.0]

  • Module changes:
    • Meta
      • Invoke-MetaModel
        • Added support for two new models: meta.llama4-maverick-17b-instruct-v1:0, meta.llama4-scout-17b-instruct-v1:0

[0.85.0] - BREAKING CHANGES

  • Module changes:
    • Converse
      • Invoke-ConverseAPI
        • parameter MediaPath renamed to ImagePath - Breaking Change
        • parameter AdditionalModelRequestField change from type PSObject to [System.Collections.Specialized.OrderedDictionary] - Breaking Change
        • Fixed bug where use of AdditionalModelRequestField was not being properly formatted
        • Added new parameters VideoPath, S3Location, and S3BucketOwner to support Vision models that can take video files
    • Updated model pricing to latest published model prices
    • Minor spelling corrections throughout

[0.75.0]

  • Module changes:
    • Mistral AI
      • Invoke-MistralAIChatModel.ps1
        • Added support for new model: mistral.pixtral-large-2502-v1:0
          • Added new MediaPath parameter as this new model supports vision
        • Fixed bug where mistral.mistral-small-2402-v1:0 was not functioning properly.

[0.70.0]

  • Module changes:
    • Amazon
      • Added new function Invoke-AmazonVideoModel to support text model: amazon.nova-reel-v1:1

[0.65.0]

  • Module changes:
    • DeepSeek
      • Added new function Invoke-DeepSeekModel to support text model: deepseek.r1-v1:0

[0.60.0]

  • Module changes:
    • Anthropic
      • Invoke-AnthropicModel
        • added support for new model: anthropic.claude-3-7-sonnet-20250219-v1:0
        • added new Thinking and ThinkingBudgetTokens parameters to support thinking and reasoning outputs for Claude 3.7 Sonnet

[0.55.0]

  • Module changes:
    • Luma
      • Added new function Invoke-LumaAIModel to support video model: luma.ray-v2:0
    • Manifest updates:
      • Require more recent version of AWS.Tools.BedrockRuntime to take advantage of new commands such as Start-BDRRAsyncInvoke
      • Require AWS.Tools.S3 to support async video file generation download from S3.
    • Minor syntax formatting corrections throughout to address new PSScriptAnalyzer rules
  • Build changes:
    • Updated bootstrap file to latest versions of modules
    • Updated all GitHub actions workflows from actions/upload-artifact@v3 to actions/upload-artifact@v4
    • pwshBedrock.build.ps1 Improvements:
      • Invoke-Formatter now runs at the end of Build step for final cleanup
      • Added new Markdown processing functionality to address two Markdown linting issues:
        • MD022/blanks-around-headings
        • MD040/fenced-code-language

[0.52.0] - BREAKING CHANGES

  • Module changes:
    • Added two new properties to all model documentation objects:
      • Image - designates if the model can generate images
      • Video - designates if the model can generate video
      • Vision - re-purposed to now designate if the model can see and describe uploaded content
        • This required some logic re-work in several functions that were previously using Vision to designate image capability
    • Added inference profile ID support to all models IAW Supported cross-region inference profiles
      • pwshBedrock will now properly format the inference profile ID for all models that support which enables the ability to make Bedrock calls to other regions
    • Reset-ModelContext - fixed issue where certain models were missing from validation.
    • Amazon
      • Invoke-AmazonNovaTextModel - added support for new Nova Text models: amazon.nova-pro-v1:0, amazon.nova-lite-v1:0, amazon.nova-micro-v1:0
      • Invoke-AmazonImageModel
        • added support for new Nova image model: amazon.nova-canvas-v1:0
        • fixed bug where files would not always save to drive
        • CfgScale parameter maximum adjusted from 10 to 9.9
    • Meta
      • Removed meta.llama2-13b-chat-v1 and meta.llama2-70b-chat-v1 from all functions as Bedrock has EOL these two models. - Breaking Change
      • Updated Invoke-MetaModel to support new model: meta.llama3-3-70b-instruct-v1:0
    • stability.ai
      • Updated Invoke-StabilityAIImageModel to support new model: stability.sd3-5-large-v1:0
    • Updated pricing for all models
  • Build changes:
    • Updated bootstrap file to latest versions of modules

[0.33.0] - BREAKING CHANGES

  • Module changes:
    • stability.ai
      • Invoke-StabilityAIDiffusionModel renamed to Invoke-StabilityAIDiffusionXLModel - Breaking Change
        • Updated help which had incorrect examples
        • Height parameter now defaults to 1024 if no user input is provided.
        • Width parameter now defaults to 1024 if no user input is provided.
      • Added Invoke-StabilityAIImageModel to support new models: stability.stable-image-core-v1:0, stability.stable-image-ultra-v1:0, stability.sd3-large-v1:0
    • AI21 Labs
      • Invoke-AI21LabsJambaModel - added support for ai21.jamba-1-5-mini-v1:0 and ai21.jamba-1-5-large-v1:0
      • Invoke-AI21LabsJurassic2Model removed. All support for ai21.j2-grande-instruct, ai21.j2-jumbo-instruct, ai21.j2-mid-v1, ai21.j2-ultra-v1 removed. These models are being EOL in Amazon Bedrock. - Breaking Change
    • Amazon
      • Invoke-AmazonTextModel - minor corrections to debug output and help
    • Cohere
      • Invoke-CohereCommandRModel - minor corrections to debug output
    • Meta
      • Invoke-MetaModel
        • Added support for new 3.2 models: meta.llama3-2-1b-instruct-v1:0, meta.llama3-2-3b-instruct-v1:0, meta.llama3-2-11b-instruct-v1:0, meta.llama3-2-90b-instruct-v1:0
        • Added vision support for new large 3.2 models
        • Added tools support for all Meta models 3.1 and higher
        • minor corrections to debug output and help
    • Mistral
      • Invoke-MistralAIModel - minor corrections to debug output and help. Adjusted Max token limit validation.
    • Anthropic
      • Invoke-AnthropicModel - added support for new models: anthropic.claude-3-5-haiku-20241022-v1:0, anthropic.claude-3-5-sonnet-20241022-v2:0
    • Updated pricing for all models
  • Build changes:
    • Updated bootstrap AWS modules from 4.1.621 to 4.1.700
    • Switched wf_Linux and wf_Windows from CodeBuild actions to GitHub Actions runners due to performance issues with CB.

[0.15.0]

  • Module changes:
    • Amazon
      • Added support for Titan Image Generator G1 V2 - amazon.titan-image-generator-v2:0
        • Added new Conditioned Image Generation parameters to Invoke-AmazonImageModel
        • Added new Color Guided Content parameters to Invoke-AmazonImageModel
        • Added new Background removal parameters to Invoke-AmazonImageModel
    • Meta
      • Added support for Llama 3.1 405B Instruct - meta.llama3-1-405b-instruct-v1:0
      • Updated pricing to reflect current Meta Llama 3.1 prices
      • Corrected Meta Llama 3.1 models to show Multilingual support as $true
    • Mistral AI
      • Updated pricing to reflect current prices
  • Build changes:
    • Updated tests to follow Pester 5 rules
    • Updated integration tests for newly added capabilities

[0.9.1]

  • Updated IconUri in manifest.

[0.9.0]

Added

  • Initial release.