@12-apps/llm 1.1.0 → 1.2.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +50 -2
- package/dist/index.d.ts +103 -4
- package/dist/index.js +263 -2
- package/dist/index.js.map +1 -1
- package/package.json +1 -1
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@12-apps/llm",
|
|
3
|
-
"version": "1.
|
|
3
|
+
"version": "1.2.0",
|
|
4
4
|
"type": "module",
|
|
5
5
|
"sideEffects": false,
|
|
6
6
|
"description": "A provider-agnostic model connector. One interface for a chat turn with tools (system blocks with cache breakpoints, strict JSON-schema tools, tool choice, output budget, effort hint) and one normalized result (text, tool calls, stop reason, usage with cache tokens). An abstract factory picks an adapter by provider id from a registry the host can extend; Anthropic, OpenAI, Gemini and xAI adapters ship over each provider's official SDK, loaded only when used. Credentials are a strategy: an API key or an OAuth access token with a host-supplied refresh. A tool-use loop reports every turn's usage. Typed errors and one host retry hook. A support ladder routes a conversation from host automation through escalating model tiers to a person, with least-privilege tools per tier (MCP tools by allow-list) bound to a host scope the model cannot set.",
|