kalibr 1.2.2__tar.gz → 1.2.4__tar.gz

kalibr-1.2.4/PKG-INFO

@@ -0,0 +1,233 @@
+Metadata-Version: 2.2
+Name: kalibr
+Version: 1.2.4
+Summary: Unified LLM Observability & Multi-Model AI Integration Framework - Deploy to GPT, Claude, Gemini, Copilot with full telemetry.
+Author-email: Kalibr Team <support@kalibr.systems>
+License: Apache-2.0
+Project-URL: Homepage, https://github.com/kalibr-ai/kalibr-sdk-python
+Project-URL: Documentation, https://kalibr.systems/docs
+Project-URL: Repository, https://github.com/kalibr-ai/kalibr-sdk-python
+Project-URL: Issues, https://github.com/kalibr-ai/kalibr-sdk-python/issues
+Keywords: ai,mcp,gpt,claude,gemini,copilot,openai,anthropic,google,microsoft,observability,telemetry,tracing,llm,schema-generation,api,multi-model,langchain,crewai
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: httpx>=0.27.0
+Requires-Dist: tiktoken>=0.8.0
+Requires-Dist: fastapi>=0.110.1
+Requires-Dist: uvicorn>=0.25.0
+Requires-Dist: pydantic>=2.6.4
+Requires-Dist: typer>=0.9.0
+Requires-Dist: python-multipart>=0.0.9
+Requires-Dist: rich>=10.0.0
+Requires-Dist: requests>=2.31.0
+Requires-Dist: opentelemetry-api>=1.20.0
+Requires-Dist: opentelemetry-sdk>=1.20.0
+Requires-Dist: opentelemetry-exporter-otlp>=1.20.0
+Provides-Extra: langchain
+Requires-Dist: langchain-core>=0.1.0; extra == "langchain"
+Provides-Extra: langchain-openai
+Requires-Dist: langchain-core>=0.1.0; extra == "langchain-openai"
+Requires-Dist: langchain-openai>=0.1.0; extra == "langchain-openai"
+Provides-Extra: langchain-anthropic
+Requires-Dist: langchain-core>=0.1.0; extra == "langchain-anthropic"
+Requires-Dist: langchain-anthropic>=0.1.0; extra == "langchain-anthropic"
+Provides-Extra: langchain-google
+Requires-Dist: langchain-core>=0.1.0; extra == "langchain-google"
+Requires-Dist: langchain-google-genai>=0.0.10; extra == "langchain-google"
+Provides-Extra: langchain-all
+Requires-Dist: langchain-core>=0.1.0; extra == "langchain-all"
+Requires-Dist: langchain-openai>=0.1.0; extra == "langchain-all"
+Requires-Dist: langchain-anthropic>=0.1.0; extra == "langchain-all"
+Requires-Dist: langchain-google-genai>=0.0.10; extra == "langchain-all"
+Provides-Extra: crewai
+Requires-Dist: crewai>=0.28.0; extra == "crewai"
+Provides-Extra: openai-agents
+Requires-Dist: openai-agents>=0.0.3; extra == "openai-agents"
+Provides-Extra: integrations
+Requires-Dist: langchain-core>=0.1.0; extra == "integrations"
+Requires-Dist: crewai>=0.28.0; extra == "integrations"
+Requires-Dist: openai-agents>=0.0.3; extra == "integrations"
+Provides-Extra: dev
+Requires-Dist: pytest>=7.4.0; extra == "dev"
+Requires-Dist: pytest-asyncio>=0.21.0; extra == "dev"
+Requires-Dist: black>=23.0.0; extra == "dev"
+Requires-Dist: ruff>=0.1.0; extra == "dev"
+
+# Kalibr SDK
+
+**Intelligent routing for AI agents.** Kalibr picks the best model for each request, learns from outcomes, and shifts traffic to what works.
+
+## Installation
+```bash
+pip install kalibr
+```
+```bash
+export KALIBR_API_KEY=kal_xxx  # Get from dashboard.kalibr.dev
+```
+
+## Quick Start
+```python
+from kalibr import Router
+
+router = Router(
+    goal="book_meeting",
+    paths=["gpt-4o", "claude-3-sonnet", "gpt-4o-mini"],
+    success_when=lambda output: "confirmed" in output.lower()
+)
+
+response = router.completion(
+    messages=[{"role": "user", "content": "Book a meeting with John tomorrow"}]
+)
+
+print(response.choices[0].message.content)
+```
+
+That's it. Kalibr handles:
+- ✅ Picking the best model (Thompson Sampling)
+- ✅ Making the API call
+- ✅ Checking success
+- ✅ Learning for next time
+- ✅ Tracing everything
+
+## How It Works
+
+1. **Define a goal** - What is your agent trying to do?
+2. **Register paths** - Which models/tools can achieve it?
+3. **Report outcomes** - Did it work?
+4. **Kalibr routes** - Traffic shifts to winners
+
+## Paths
+
+A path is a model + optional tools + optional params:
+```python
+# Simple: just models
+paths = ["gpt-4o", "claude-3-sonnet"]
+
+# With tools
+paths = [
+    {"model": "gpt-4o", "tools": ["web_search"]},
+    {"model": "claude-3-sonnet", "tools": ["web_search", "browser"]},
+]
+
+# With params
+paths = [
+    {"model": "gpt-4o", "params": {"temperature": 0.7}},
+    {"model": "gpt-4o", "params": {"temperature": 0.2}},
+]
+```
+
+## Success Criteria
+
+### Auto-detect from output
+```python
+router = Router(
+    goal="summarize",
+    paths=["gpt-4o", "claude-3-sonnet"],
+    success_when=lambda output: len(output) > 100
+)
+```
+
+### Manual reporting
+```python
+router = Router(goal="book_meeting", paths=["gpt-4o", "claude-3-sonnet"])
+
+response = router.completion(messages=[...])
+
+# Your verification logic
+meeting_created = check_calendar_api()
+
+router.report(success=meeting_created)
+```
+
+## Framework Integration
+
+### LangChain
+```python
+from kalibr import Router
+
+router = Router(goal="summarize", paths=["gpt-4o", "claude-3-sonnet"])
+llm = router.as_langchain()
+
+chain = prompt | llm | parser
+result = chain.invoke({"text": "..."})
+```
+
+### CrewAI
+```python
+from kalibr import Router
+
+router = Router(goal="research", paths=["gpt-4o", "claude-3-sonnet"])
+
+agent = Agent(
+    role="Researcher",
+    llm=router.as_langchain(),
+    ...
+)
+```
+
+## Observability (Included)
+
+Every call is automatically traced:
+
+- Token counts and costs
+- Latency (p50, p95, p99)
+- Tool usage
+- Errors with stack traces
+
+View in the [dashboard](https://dashboard.kalibr.dev) or use callback handlers directly:
+```python
+from kalibr_langchain import KalibrCallbackHandler
+
+handler = KalibrCallbackHandler()
+chain.invoke({"input": "..."}, config={"callbacks": [handler]})
+```
+
+## Pricing
+
+| Tier | Routing Decisions | Price |
+|------|-------------------|-------|
+| Free | 1,000/month | $0 |
+| Pro | 50,000/month | $49/month |
+| Enterprise | Unlimited | Custom |
+
+## API Reference
+
+### Router
+```python
+Router(
+    goal: str,                    # Required: name of the goal
+    paths: List[str | dict],      # Models/tools to route between
+    success_when: Callable,       # Optional: auto-evaluate success
+    exploration_rate: float,      # Optional: 0.0-1.0, default 0.1
+)
+```
+
+### Methods
+```python
+router.completion(messages, **kwargs)  # Make routed request
+router.report(success, reason=None)    # Report outcome manually
+router.add_path(model, tools=None)     # Add path dynamically
+router.as_langchain()                  # Get LangChain-compatible LLM
+```
+
+## Links
+
+- [Documentation](https://docs.kalibr.dev)
+- [Dashboard](https://dashboard.kalibr.dev)
+- [GitHub](https://github.com/kalibr-ai/kalibr-sdk-python)
+
+## License
+
+MIT

kalibr-1.2.4/README.md

@@ -0,0 +1,166 @@
+# Kalibr SDK
+
+**Intelligent routing for AI agents.** Kalibr picks the best model for each request, learns from outcomes, and shifts traffic to what works.
+
+## Installation
+```bash
+pip install kalibr
+```
+```bash
+export KALIBR_API_KEY=kal_xxx  # Get from dashboard.kalibr.dev
+```
+
+## Quick Start
+```python
+from kalibr import Router
+
+router = Router(
+    goal="book_meeting",
+    paths=["gpt-4o", "claude-3-sonnet", "gpt-4o-mini"],
+    success_when=lambda output: "confirmed" in output.lower()
+)
+
+response = router.completion(
+    messages=[{"role": "user", "content": "Book a meeting with John tomorrow"}]
+)
+
+print(response.choices[0].message.content)
+```
+
+That's it. Kalibr handles:
+- ✅ Picking the best model (Thompson Sampling)
+- ✅ Making the API call
+- ✅ Checking success
+- ✅ Learning for next time
+- ✅ Tracing everything
+
+## How It Works
+
+1. **Define a goal** - What is your agent trying to do?
+2. **Register paths** - Which models/tools can achieve it?
+3. **Report outcomes** - Did it work?
+4. **Kalibr routes** - Traffic shifts to winners
+
+## Paths
+
+A path is a model + optional tools + optional params:
+```python
+# Simple: just models
+paths = ["gpt-4o", "claude-3-sonnet"]
+
+# With tools
+paths = [
+    {"model": "gpt-4o", "tools": ["web_search"]},
+    {"model": "claude-3-sonnet", "tools": ["web_search", "browser"]},
+]
+
+# With params
+paths = [
+    {"model": "gpt-4o", "params": {"temperature": 0.7}},
+    {"model": "gpt-4o", "params": {"temperature": 0.2}},
+]
+```
+
+## Success Criteria
+
+### Auto-detect from output
+```python
+router = Router(
+    goal="summarize",
+    paths=["gpt-4o", "claude-3-sonnet"],
+    success_when=lambda output: len(output) > 100
+)
+```
+
+### Manual reporting
+```python
+router = Router(goal="book_meeting", paths=["gpt-4o", "claude-3-sonnet"])
+
+response = router.completion(messages=[...])
+
+# Your verification logic
+meeting_created = check_calendar_api()
+
+router.report(success=meeting_created)
+```
+
+## Framework Integration
+
+### LangChain
+```python
+from kalibr import Router
+
+router = Router(goal="summarize", paths=["gpt-4o", "claude-3-sonnet"])
+llm = router.as_langchain()
+
+chain = prompt | llm | parser
+result = chain.invoke({"text": "..."})
+```
+
+### CrewAI
+```python
+from kalibr import Router
+
+router = Router(goal="research", paths=["gpt-4o", "claude-3-sonnet"])
+
+agent = Agent(
+    role="Researcher",
+    llm=router.as_langchain(),
+    ...
+)
+```
+
+## Observability (Included)
+
+Every call is automatically traced:
+
+- Token counts and costs
+- Latency (p50, p95, p99)
+- Tool usage
+- Errors with stack traces
+
+View in the [dashboard](https://dashboard.kalibr.dev) or use callback handlers directly:
+```python
+from kalibr_langchain import KalibrCallbackHandler
+
+handler = KalibrCallbackHandler()
+chain.invoke({"input": "..."}, config={"callbacks": [handler]})
+```
+
+## Pricing
+
+| Tier | Routing Decisions | Price |
+|------|-------------------|-------|
+| Free | 1,000/month | $0 |
+| Pro | 50,000/month | $49/month |
+| Enterprise | Unlimited | Custom |
+
+## API Reference
+
+### Router
+```python
+Router(
+    goal: str,                    # Required: name of the goal
+    paths: List[str | dict],      # Models/tools to route between
+    success_when: Callable,       # Optional: auto-evaluate success
+    exploration_rate: float,      # Optional: 0.0-1.0, default 0.1
+)
+```
+
+### Methods
+```python
+router.completion(messages, **kwargs)  # Make routed request
+router.report(success, reason=None)    # Report outcome manually
+router.add_path(model, tools=None)     # Add path dynamically
+router.as_langchain()                  # Get LangChain-compatible LLM
+```
+
+## Links
+
+- [Documentation](https://docs.kalibr.dev)
+- [Dashboard](https://dashboard.kalibr.dev)
+- [GitHub](https://github.com/kalibr-ai/kalibr-sdk-python)
+
+## License
+
+MIT

{kalibr-1.2.2 → kalibr-1.2.4}/kalibr/__init__.py

@@ -92,6 +92,7 @@ from .intelligence import (
     register_path,
     decide,
 )
+from .router import Router
 
 if os.getenv("KALIBR_AUTO_INSTRUMENT", "true").lower() == "true":
     # Setup OpenTelemetry collector
@@ -163,4 +164,5 @@ __all__ = [
     "get_recommendation",
     "register_path",
     "decide",
+    "Router",
 ]