llmist 1.2.0 → 1.3.1

package/README.md

@@ -1,11 +1,11 @@
 # llmist
 
 [![CI](https://github.com/zbigniewsobiecki/llmist/actions/workflows/ci.yml/badge.svg)](https://github.com/zbigniewsobiecki/llmist/actions/workflows/ci.yml)
-[![codecov](https://codecov.io/gh/zbigniewsobiecki/llmist/graph/badge.svg)](https://codecov.io/gh/zbigniewsobiecki/llmist)
+[![codecov](https://codecov.io/gh/zbigniewsobiecki/llmist/graph/badge.svg?branch=dev)](https://codecov.io/gh/zbigniewsobiecki/llmist)
 [![npm version](https://img.shields.io/npm/v/llmist.svg)](https://www.npmjs.com/package/llmist)
 [![License](https://img.shields.io/badge/License-MIT-blue.svg)](https://opensource.org/licenses/MIT)
 
-> **Universal TypeScript LLM client with streaming-first tool execution and simple, extensible agent framework**
+> **Universal TypeScript LLM client with own function calling grammar, streaming-first tool execution and simple, extensible agent framework**
 
 > **⚠️ EARLY WORK IN PROGRESS** - This library is under active development. APIs may change without notice. Use in production at your own risk.
 
@@ -15,12 +15,13 @@ llmist is an asynchonous, streaming-first, provider-agnostic LLM client that mak
 
 ## 🎯 Why llmist?
 
-- **🌍 Universal** - Works with any LLM provider (OpenAI, Anthropic, Gemini, custom)
-- **📝 No Structured Outputs** - Flexible TOML/YAML/JSON grammar works with any text model
-- **⚡ Streaming-First** - Built for real-time responses and efficient error handling
-- **🪝 Powerful Hooks** - Monitor, customize, and control every step of execution
-- **🎨 Beautiful API** - Fluent builder pattern with model shortcuts and presets
-- **🧪 Testing-Friendly** - Built-in mocking system for zero-cost testing
+- **🌍 Universal** - Works with [any LLM provider](./docs/PROVIDERS.md) (OpenAI, Anthropic, Gemini, [custom](./docs/CUSTOM_MODELS.md)) and easy to integrate more
+- **📝 No Structured Outputs** - Simple [block format](./docs/BLOCK_FORMAT.md) with streaming oriented tool calling
+- **🪝 Powerful Hooks** - Monitor, customize, and control [every step of execution](./docs/HOOKS.md)
+- **🎨 Fluent API** - [Builder pattern](./docs/CONFIGURATION.md) with model shortcuts and presets
+- **🧪 Testing-Friendly** - Built-in [mocking system](./docs/TESTING.md) for zero-cost testing
+- **⌨️ Convenient CLI** - [Capable CLI](./docs/CLI.md) showcasing how to build on top of llmist
+- **🏦 Cost-aware** - [Cost APIs per model](./docs/MODEL_CATALOG.md) + prompt-caching aware accounting
 
 ---
 
@@ -49,18 +50,6 @@ bunx llmist agent "Calculate 15 * 23" --gadget ./calculator.ts --model sonnet
 cat document.txt | llmist complete "Summarize" --model gpt-5-nano
 ```
 
-**Built-in gadgets** are included by default in agent mode:
-- `AskUser` - Prompts for user input when clarification is needed
-- `TellUser` - Displays important messages (info/success/warning/error) and can end conversations
-
-```bash
-# Disable built-in gadgets
-bunx llmist agent "Task" --no-builtins -g ./my-tools.ts
-
-# Override parameter format (default: toml, options: toml, yaml, json, auto)
-bunx llmist agent "Task" --parameter-format yaml -g ./my-tools.ts
-```
-
 📖 **[CLI Reference](./docs/CLI.md)** | **[CLI Gadgets Guide](./docs/CLI_GADGETS.md)**
 
 
@@ -100,7 +89,7 @@ const answer = await LLMist.createAgent()
 console.log(answer); // "15 times 23 equals 345"
 ```
 
-**That's it!** N
+**That's it!**
 
 📖 **[Getting Started Guide](./docs/GETTING_STARTED.md)** - Learn more in 5 minutes
 
@@ -354,6 +343,7 @@ Uses provider-specific methods (tiktoken for OpenAI, native APIs for Anthropic/G
 
 **Core Concepts**
 - **[Gadgets (Tools)](./docs/GADGETS.md)** - Creating custom functions
+- **[Block Format](./docs/BLOCK_FORMAT.md)** - Parameter syntax reference
 - **[Hooks](./docs/HOOKS.md)** - Lifecycle monitoring and control
 - **[Streaming](./docs/STREAMING.md)** - Real-time response handling
 - **[Human-in-the-Loop](./docs/HUMAN_IN_LOOP.md)** - Interactive workflows
@@ -433,60 +423,7 @@ bun run format
 
 ## 🤝 Contributing
 
-Contributions welcome! Please ensure:
-
-1. ✅ All tests pass: `bun test`
-2. ✅ Code is formatted: `bun run format`
-3. ✅ Linting passes: `bun run lint`
-4. ✅ Types are properly defined
-5. ✅ Examples/docs updated for API changes
-
-### Commit Message Convention
-
-This project follows [Conventional Commits](https://www.conventionalcommits.org/) specification. All commit messages must be formatted as:
-
-```
-<type>(<scope>): <subject>
-```
-
-**Types:**
-- `feat:` - New feature (triggers minor version bump)
-- `fix:` - Bug fix (triggers patch version bump)
-- `docs:` - Documentation only changes
-- `style:` - Code style changes (formatting, missing semi-colons, etc)
-- `refactor:` - Code refactoring without feature changes
-- `perf:` - Performance improvements
-- `test:` - Adding or updating tests
-- `build:` - Build system or dependency changes
-- `ci:` - CI configuration changes
-- `chore:` - Other changes that don't modify src or test files
-
-**Breaking Changes:** Add `BREAKING CHANGE:` in the footer to trigger major version bump.
-
-**Examples:**
-```bash
-feat(agent): add support for streaming tool calls
-fix(cli): prevent crash on invalid gadget path
-docs: update API documentation for v2
-```
-
-**Note:** Git hooks will validate your commit messages locally.
-
-### Release Process
-
-Releases are fully automated using [semantic-release](https://github.com/semantic-release/semantic-release):
-
-1. Merge PR to `main` branch
-2. CI workflow runs automatically
-3. If CI passes, release workflow:
-   - Analyzes commits since last release
-   - Determines version bump based on commit types
-   - Updates `package.json` and `CHANGELOG.md`
-   - Creates git tag and GitHub release
-   - Publishes to npm
-   - Syncs changes back to `dev` branch
-
-**No manual version bumps needed!**
+Contributions welcome! See **[CONTRIBUTING.md](./CONTRIBUTING.md)** for guidelines, commit conventions, and release process.
 
 ---