docs: XaiLLM, NovitaLLM, WriterLLM, WeatherTool, StripeTool, NewsTool, HTMLTextSplitter

docs/agents/tools.md

@@ -1334,3 +1334,93 @@ The `whatsapp:` prefix is handled automatically — pass a plain E.164 number fo
 - Tier-based rate limits (1K/10K/100K unique users per 24h) apply to business-initiated conversations only
 - Business-initiated conversations (marketing, utility, authentication) carry an additional per-conversation charge on top of the per-message fee; user-initiated conversations have no per-conversation fee
 :::
+
+## NewsTool
+
+Fetch news headlines and search articles via NewsAPI. No extra dependencies — stdlib `urllib` only.
+
+```python
+from synapsekit import NewsTool
+
+tool = NewsTool(api_key="your-newsapi-key")
+# or set NEWS_API_KEY env var
+
+# Get top headlines
+result = await tool.run(action="get_headlines", country="us", category="technology")
+
+# Search articles
+result = await tool.run(action="search", query="AI developments", sort_by="publishedAt")
+
+print(result.output)  # JSON array of articles
+```
+
+| Parameter | Default | Description |
+|---|---|---|
+| `action` | — | `get_headlines` or `search` (required) |
+| `query` | — | Search query (required for `search`) |
+| `country` | `us` | 2-letter country code for headlines |
+| `category` | — | `business`, `entertainment`, `general`, `health`, `science`, `sports`, `technology` |
+| `from_date` | — | Oldest article date (ISO 8601) |
+| `sort_by` | — | `relevancy`, `popularity`, or `publishedAt` |
+
+Auth via constructor arg or `NEWS_API_KEY` env var. Get a free API key at [newsapi.org](https://newsapi.org).
+
+## WeatherTool
+
+Get current weather and short-term forecasts via OpenWeatherMap. No extra dependencies — stdlib `urllib` only.
+
+```python
+from synapsekit import WeatherTool
+
+tool = WeatherTool()
+# set OPENWEATHERMAP_API_KEY env var
+
+# Current weather
+result = await tool.run(action="current", city="London,UK")
+
+# 3-day forecast
+result = await tool.run(action="forecast", city="Tokyo", days=3)
+
+print(result.output)
+```
+
+| Parameter | Default | Description |
+|---|---|---|
+| `action` | `current` | `current` or `forecast` |
+| `city` | — | City name, e.g. `Delhi` or `London,UK` (required) |
+| `days` | `5` | Forecast days 1–5 (for `forecast` action) |
+
+Auth via `OPENWEATHERMAP_API_KEY` env var.
+
+## StripeTool
+
+Read-only Stripe data lookup. No extra dependencies — stdlib `urllib` only.
+
+```python
+from synapsekit import StripeTool
+
+tool = StripeTool()
+# set STRIPE_API_KEY env var
+
+# Look up customer
+result = await tool.run(action="get_customer", id_or_email="cus_123")
+
+# List invoices
+result = await tool.run(action="list_invoices", customer_id="cus_123", limit=5)
+
+# Get charge details
+result = await tool.run(action="get_charge", charge_id="ch_abc")
+
+# List products with pricing
+result = await tool.run(action="list_products", limit=10)
+```
+
+| Parameter | Default | Description |
+|---|---|---|
+| `action` | — | `get_customer`, `list_invoices`, `get_charge`, `list_products` (required) |
+| `id_or_email` | — | Customer ID (`cus_...`) or email for `get_customer` |
+| `customer_id` | — | Stripe customer ID for `list_invoices` |
+| `charge_id` | — | Stripe charge ID for `get_charge` |
+| `limit` | `10` | Max results for list endpoints (max 100) |
+
+Auth via `STRIPE_API_KEY` env var (use restricted read-only keys in production).

docs/intro.md

@@ -4,7 +4,7 @@ sidebar_position: 1
 
 # Introduction
 
-**SynapseKit** is an async-native Python framework for building LLM applications — RAG pipelines, tool-using agents, and graph workflows. Streaming-first, transparent API, 2 hard deps. 27 providers · 43 tools · 26 loaders · 9 vector stores.
+**SynapseKit** is an async-native Python framework for building LLM applications — RAG pipelines, tool-using agents, and graph workflows. Streaming-first, transparent API, 2 hard deps. 30 providers · 45 tools · 26 loaders · 9 vector stores.
 
 It is designed from the ground up to be **async-native** and **streaming-first**. Every public API is `async`. Streaming tokens is the default, not an opt-in. There are no hidden chains, no magic callbacks, no global state.
 
@@ -65,7 +65,7 @@ InMemoryVectorStore (built-in, `.npz` persistence), ChromaDB, FAISS, Qdrant, Pin
 `ReActAgent` — Thought → Action → Observation loop, works with any LLM.
 `FunctionCallingAgent` — native `tool_calls` / `tool_use` for OpenAI, Anthropic, Gemini, and Mistral.
 `AgentExecutor` — unified runner, picks the right agent from config.
-43 built-in tools: Calculator, PythonREPL, FileRead, FileWrite, FileList, WebSearch, DuckDuckGoSearch, SQL, HTTP, GraphQL, DateTime, Regex, JSONQuery, HumanInput, Wikipedia, Summarization, SentimentAnalysis, Translation, WebScraper, Shell, SQLSchemaInspection, PDFReader, ArxivSearch, TavilySearch, Email, GitHubAPI, PubMedSearch, VectorSearch, YouTubeSearch, Slack, Notion, Jira, BraveSearch, APIBuilder, GoogleCalendar, AWSLambda, ImageAnalysis, TextToSpeech, SpeechToText, BingSearch, WolframAlpha, GoogleSearch, Twilio.
+45 built-in tools: Calculator, PythonREPL, FileRead, FileWrite, FileList, WebSearch, DuckDuckGoSearch, SQL, HTTP, GraphQL, DateTime, Regex, JSONQuery, HumanInput, Wikipedia, Summarization, SentimentAnalysis, Translation, WebScraper, Shell, SQLSchemaInspection, PDFReader, ArxivSearch, TavilySearch, Email, GitHubAPI, PubMedSearch, VectorSearch, YouTubeSearch, Slack, Notion, Jira, BraveSearch, APIBuilder, GoogleCalendar, AWSLambda, ImageAnalysis, TextToSpeech, SpeechToText, BingSearch, WolframAlpha, GoogleSearch, Twilio, NewsTool, WeatherTool, StripeTool.
 
 → [Agent docs](/docs/agents/overview)
 

docs/llms/novita.md

@@ -0,0 +1,48 @@
+---
+sidebar_position: 29
+---
+
+# NovitaAI
+
+NovitaAI hosts popular open models (Llama, Mistral, Qwen, etc.) via an OpenAI-compatible API.
+
+## Install
+
+```bash
+pip install synapsekit[openai]
+```
+
+## Usage
+
+```python
+from synapsekit.llm.novita import NovitaLLM
+from synapsekit import LLMConfig
+
+config = LLMConfig(
+    model="meta-llama/llama-3.1-8b-instruct",
+    api_key="nvt_...",
+    provider="novita",
+)
+
+llm = NovitaLLM(config)
+
+# Streaming
+async for token in llm.stream("Summarize the French Revolution"):
+    print(token, end="")
+
+# Generate
+response = await llm.generate("Explain async/await in Python")
+```
+
+## Available models
+
+NovitaAI hosts hundreds of models. Popular choices:
+
+| Model | Notes |
+|---|---|
+| `meta-llama/llama-3.1-8b-instruct` | Fast, 8B params |
+| `meta-llama/llama-3.1-70b-instruct` | High quality |
+| `mistralai/mistral-7b-instruct-v0.3` | Efficient |
+| `qwen/qwen2.5-72b-instruct` | Strong multilingual |
+
+See the full list at [novita.ai/models](https://novita.ai/models/text-generation).

docs/llms/overview.md

@@ -79,6 +79,9 @@ See [Caching & Retries](/docs/llms/caching-retries) for details on response cach
 | Aleph Alpha | `AlephAlphaLLM` | `pip install synapsekit[aleph-alpha]` | `"aleph-alpha"` |
 | Hugging Face | `HuggingFaceLLM` | `pip install synapsekit[huggingface]` | `"huggingface"` |
 | SambaNova | `SambaNovaLLM` | `pip install synapsekit[openai]` | `"sambanova"` |
+| xAI (Grok) | `XaiLLM` | `pip install synapsekit[openai]` | `"xai"` |
+| NovitaAI | `NovitaLLM` | `pip install synapsekit[openai]` | `"novita"` |
+| Writer (Palmyra) | `WriterLLM` | `pip install synapsekit[openai]` | `"writer"` |
 
 ## Auto-detection
 
@@ -138,6 +141,9 @@ The `TokenTracer` in `RAGPipeline` aggregates this across all calls.
 - [Aleph Alpha](./aleph-alpha) — European LLMs, German-language and multilingual
 - [Hugging Face](./huggingface) — thousands of open-source models via Inference API and Dedicated Endpoints
 - [SambaNova](./sambanova) — fast inference on Llama, Qwen, and other open models
+- [xAI](./xai) — Grok models (grok-2, grok-2-mini)
+- [NovitaAI](./novita) — hosted open models (Llama, Mistral, Qwen, etc.)
+- [Writer](./writer) — Palmyra models including domain-specific (palmyra-med, palmyra-fin)
 - [Caching & Retries](./caching-retries) — LRU caching, exponential backoff, rate limiting
 - [CostRouter & FallbackChain](./cost-router) — route to cheapest model or cascade on failure
 - [Cost Tracker](../observability/cost-tracker) — attribute and budget LLM spending

docs/llms/writer.md

@@ -0,0 +1,59 @@
+---
+sidebar_position: 30
+---
+
+# Writer (Palmyra)
+
+Writer's Palmyra models via the OpenAI-compatible API. Includes domain-specific models for medicine and finance.
+
+## Install
+
+```bash
+pip install synapsekit[openai]
+```
+
+## Usage
+
+```python
+from synapsekit.llm.writer import WriterLLM
+from synapsekit import LLMConfig
+
+config = LLMConfig(
+    model="palmyra-x-004",
+    api_key="...",
+    provider="writer",
+)
+
+llm = WriterLLM(config)
+
+# Streaming
+async for token in llm.stream("Draft a contract summary"):
+    print(token, end="")
+
+# Generate
+response = await llm.generate("Explain HIPAA compliance")
+```
+
+## Available models
+
+| Model | Notes |
+|---|---|
+| `palmyra-x-004` | Latest general purpose |
+| `palmyra-x-003-instruct` | Instruction-tuned |
+| `palmyra-med` | Medical domain |
+| `palmyra-fin` | Financial domain |
+
+## Function calling
+
+```python
+from synapsekit import AgentExecutor, AgentConfig, CalculatorTool
+
+config = AgentConfig(
+    model="palmyra-x-004",
+    api_key="...",
+    provider="writer",
+    tools=[CalculatorTool()],
+)
+executor = AgentExecutor(config)
+result = await executor.run("Calculate compound interest on $10,000 at 5% for 3 years")
+```

docs/llms/xai.md

@@ -0,0 +1,58 @@
+---
+sidebar_position: 28
+---
+
+# xAI (Grok)
+
+xAI's Grok models via the OpenAI-compatible API.
+
+## Install
+
+```bash
+pip install synapsekit[openai]
+```
+
+## Usage
+
+```python
+from synapsekit.llm.xai import XaiLLM
+from synapsekit import LLMConfig
+
+config = LLMConfig(
+    model="grok-2",
+    api_key="xai-...",
+    provider="xai",
+)
+
+llm = XaiLLM(config)
+
+# Streaming
+async for token in llm.stream("Explain the trolley problem"):
+    print(token, end="")
+
+# Generate
+response = await llm.generate("What is entropy?")
+```
+
+## Available models
+
+| Model | Notes |
+|---|---|
+| `grok-beta` | Original Grok |
+| `grok-2` | Latest generation |
+| `grok-2-mini` | Faster, lower cost |
+
+## Function calling
+
+```python
+from synapsekit import AgentExecutor, AgentConfig, CalculatorTool
+
+config = AgentConfig(
+    model="grok-2",
+    api_key="xai-...",
+    provider="xai",
+    tools=[CalculatorTool()],
+)
+executor = AgentExecutor(config)
+result = await executor.run("What is 15% of 284?")
+```

docs/rag/splitter.md

@@ -297,3 +297,24 @@ splitter = SentenceSplitter()
 chunks = splitter.split("First sentence. Second sentence. Third sentence.")
 # ["First sentence.", "Second sentence.", "Third sentence."]
 ```
+
+## HTMLTextSplitter
+
+Splits HTML documents on block-level tags and returns plain-text chunks. Uses stdlib `html.parser` — no extra dependencies.
+
+```python
+from synapsekit import HTMLTextSplitter
+
+splitter = HTMLTextSplitter(chunk_size=512, chunk_overlap=50)
+chunks = splitter.split("<h1>Title</h1><p>First paragraph.</p><p>Second paragraph.</p>")
+# ["Title", "First paragraph.", "Second paragraph."]
+```
+
+Block tags used as split boundaries: `h1`, `h2`, `h3`, `h4`, `h5`, `h6`, `p`, `div`, `section`, `article`, `li`, `blockquote`, `pre`.
+
+Long sections that exceed `chunk_size` fall back to `RecursiveCharacterTextSplitter`.
+
+| Parameter | Default | Description |
+|---|---|---|
+| `chunk_size` | `512` | Maximum characters per chunk |
+| `chunk_overlap` | `50` | Characters of overlap between consecutive chunks |