Skip to content

About

lm15 for .NET: an early port, written against the lm15 contract of 2026-09-11; not published.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Repository files navigation

lm15 for .NET (early port)

lm15 is one request and response model for every major AI model provider: write a request once and send it to OpenAI, Anthropic, Gemini, xAI, Groq, DeepSeek, OpenRouter, a cloud or a local model, by changing the model string. Each language implements it separately and is graded by one shared contract.

Early port, incomplete, not published. Written against the lm15 contract of 2026-09-11 (cfed007, in CONTRACT_PIN). At that commit it passed the serialization (115), token (43) and credential (43) directions; the provider dialects, streaming and the other surfaces were not written. It has not been updated since.

lm15 is released and stable (2.0) in Python, TypeScript, Rust, Go and R. Guides: lm15.dev. For production, use one of those.

What's here

.NET 8, zero runtime dependencies: every canonical type and its exact JSON, the error hierarchy, credentials (policies, stores, cloud chains, OAuth, the doctor), SigV4 and RS256 signing, compat presets, the provider registry, transport and SSE. Not yet: the provider dialects, stream assembly and the surfaces that use them. PORTING.md is the map for whoever continues it.

Try it

dotnet build -c Release
dotnet test

Conformance

python3 harness/check.py --shim dotnet --direction serde --no-check-pin   # from lm15-contract, with the shim registered

The full guide written with this port, with every API and its stated differences from the Python reference: docs/guide.md.

About

lm15 for .NET: an early port, written against the lm15 contract of 2026-09-11; not published.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages