Hey everyone! It's been a while since we announce...

# http4k

[dave]

Hey everyone!

It's been a while since we announced any changes to http4k-connect on this channel, so we thought we'd give you all a summary of what we've been up to. We've just released *v*5.18.0.0.

Recent versions have included the following.

- 🚀 AI connectors - we recognise that lots of teams are wanting to play with the latest cool technologies, but we're also pretty saddened by the lack of observability options in the various LLM Client SDKs and tool libraries. As such, we've decided to focus on building adapters to help teams developing for these shiny new toys which will slide right into your http4k stack and allow total observability with our consistent HTTP client interface. We're especially concerned with creating Fakes for these AI services wherever possible due to the environmental impact of using the real thing with CI pipelines and local test suites. Obviously the temperature of the fakes is also 0, so that makes tests using them completely deterministic 🙂 :
- There is now a unified http4k-connect-ai module prefix for all AI connector modules
- http4k-connect-ai-openai : Module to interact with OpenAI, including a FakeOpenAI which generates text and image responses, as well as embeddings.
- http4k-connect-ai-lmstudio : Module to interact with the locally running LmStudio LLM runner, including a FakeLmStudio which generates text responses, as well as embeddings.
- http4k-connect-ai-ollama : Module to interact with the locally running Ollama LLM runner, including a FakeOllama which generates text responses, as well as embeddings.
- http4k-connect-ai-langchain : A set of Chat, Image and Embedding models which use the above LLM adapter modulesto integrate directly with LangChain4j. Expect these to be added to as the set of LLM adapters grows
- If you're using an LLM that isn't listed above, we'd also very much welcome contributions for adapters to others (anthropic, gemini etc..)
- 🚀 L*ightweight AWS connectors:*
- http4k-connect-amazon-evidently - Feature flagging adapter and fake
- `http4k-connect-amazon-evidently`(- Serverless app container control API adapter and fake
- http4k-connect-amazon-sqs - Support for the new JSON APIs for SQS

One final thing, be sure to check out @Andrew O'Hara’s great KotlinConf

talk▾

about his adventures optimising the cold starts of AWS Lambdas

[Andrew O'Hara]

- `http4k-connect-amazon-evidently`(- Serverless app container control API adapter and fake

Wrong module name? Interested to look into it.

[dave]

Oops - that should be apprunner!

[dave]

It's quite basic - we use it for service discovery inside tooling, but the base is there

[dave]

Unfortunately I can't edit the message anymore 🙄 . Full list of actions at https://docs.aws.amazon.com/apprunner/latest/api/API_Operations.html