PyStreamMCP 2.1.0__tar.gz → 3.2.0__tar.gz

pystreammcp-3.2.0/LICENSE

@@ -0,0 +1,22 @@
+Proprietary License — Free to use with explicit attribution
+
+Copyright © 2026 Georgi Mammen Mullassery
+
+Permission is granted to any person obtaining a copy of this software to use,
+copy, modify, and distribute this software and its documentation, provided that:
+
+1. The above copyright notice and this permission notice appear in all copies
+   or substantial portions of the software.
+
+2. Explicit attribution to the original author (Georgi Mammen Mullassery) is
+   provided in any project documentation and significant code comments where
+   this software is used.
+
+3. The software is provided "AS IS" without warranty of any kind, express or
+   implied, including but not limited to the warranties of merchantability,
+   fitness for a particular purpose, or non-infringement.
+
+4. In no event shall the author be liable for any claim, damages, or other
+   liability arising from the use of this software.
+
+For more information, visit: https://github.com/Mullassery

pystreammcp-3.2.0/PKG-INFO

@@ -0,0 +1,231 @@
+Metadata-Version: 2.4
+Name: PyStreamMCP
+Version: 3.2.0
+Summary: Intelligent MCP orchestration hub - Intent understanding, capability matching, tool ranking with cohesive foundation
+Author-email: Georgi Mammen Mullassery <mullassery@gmail.com>
+License: Proprietary License — Free to use with explicit attribution
+Project-URL: Repository, https://github.com/Mullassery/PyStreamMCP
+Project-URL: Issues, https://github.com/Mullassery/PyStreamMCP/issues
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: Other/Proprietary License
+Classifier: Programming Language :: Python :: 3
+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: Programming Language :: Python :: 3.13
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: pydantic>=2.0
+Requires-Dist: python-frontmatter>=1.0.0
+Requires-Dist: click>=8.0
+Requires-Dist: rich>=13.0
+Requires-Dist: httpx>=0.24.0
+Requires-Dist: toml>=0.10; python_version < "3.11"
+Provides-Extra: dev
+Requires-Dist: pytest>=7.0; extra == "dev"
+Requires-Dist: pytest-cov>=4.0; extra == "dev"
+Requires-Dist: pytest-asyncio>=0.21.0; extra == "dev"
+Provides-Extra: api
+Requires-Dist: fastapi>=0.104.0; extra == "api"
+Requires-Dist: uvicorn[standard]>=0.24.0; extra == "api"
+Requires-Dist: pydantic-settings>=2.0; extra == "api"
+Requires-Dist: flask>=2.3.0; extra == "api"
+Provides-Extra: mcp
+Requires-Dist: mcp>=0.1.0; extra == "mcp"
+Provides-Extra: langchain
+Requires-Dist: langchain>=0.1.0; extra == "langchain"
+Provides-Extra: llamaindex
+Requires-Dist: llama-index>=0.9.0; extra == "llamaindex"
+Provides-Extra: semantic-kernel
+Requires-Dist: semantic-kernel>=0.4.0; extra == "semantic-kernel"
+Provides-Extra: crewai
+Requires-Dist: crewai>=0.1.0; extra == "crewai"
+Provides-Extra: pydantic-ai
+Requires-Dist: pydantic-ai>=0.1.0; extra == "pydantic-ai"
+Provides-Extra: haystack
+Requires-Dist: haystack-ai>=2.0.0; extra == "haystack"
+Provides-Extra: temporal
+Requires-Dist: temporalio>=1.0.0; extra == "temporal"
+Provides-Extra: airflow
+Requires-Dist: apache-airflow>=2.7.0; extra == "airflow"
+Provides-Extra: all-integrations
+Requires-Dist: langchain>=0.1.0; extra == "all-integrations"
+Requires-Dist: llama-index>=0.9.0; extra == "all-integrations"
+Requires-Dist: semantic-kernel>=0.4.0; extra == "all-integrations"
+Requires-Dist: crewai>=0.1.0; extra == "all-integrations"
+Requires-Dist: pydantic-ai>=0.1.0; extra == "all-integrations"
+Requires-Dist: haystack-ai>=2.0.0; extra == "all-integrations"
+Requires-Dist: temporalio>=1.0.0; extra == "all-integrations"
+Requires-Dist: apache-airflow>=2.7.0; extra == "all-integrations"
+Dynamic: license-file
+
+# PyStreamMCP
+
+[![CI](https://github.com/Mullassery/PyStreamMCP/actions/workflows/ci.yml/badge.svg)](https://github.com/Mullassery/PyStreamMCP/actions/workflows/ci.yml)
+
+An AI-agent intelligence layer: query planning, context discovery, and
+token-cost optimization, exposed both as a Python SDK and as an MCP
+(Model Context Protocol) tool server, plus a small event-driven
+orchestration/federation layer for coordinating multiple MCP endpoints.
+
+Pure Python — `pip install` just works, no Rust toolchain or compiler
+required.
+
+## Install
+
+```bash
+pip install PyStreamMCP
+```
+
+Optional extras, install what you actually need:
+
+```bash
+pip install "PyStreamMCP[api]"          # FastAPI + Flask HTTP servers
+pip install "PyStreamMCP[mcp]"          # MCP protocol client library
+pip install "PyStreamMCP[langchain]"    # LangChain adapter
+pip install "PyStreamMCP[llamaindex]"   # LlamaIndex adapter
+pip install "PyStreamMCP[semantic-kernel]"  # Semantic Kernel adapter
+pip install "PyStreamMCP[all-integrations]" # every framework adapter
+```
+
+Requires Python 3.9+.
+
+## Quick start
+
+```python
+from pystreammcp import Agent
+
+agent = Agent(
+    agent_id="recommendation_engine",
+    name="Product Recommendation Engine",
+    optimization_strategy="token_efficient",  # or "balanced" / "quality_first"
+    max_tokens=1000,
+)
+
+result = agent.query("What are the top 10 customers by lifetime value?")
+
+print(f"Baseline tokens:  {result.baseline_tokens}")
+print(f"Optimized tokens: {result.optimized_tokens}")
+print(f"Reduction:        {result.cost_reduction_percent:.1f}%")
+```
+
+Token counts are estimated from the actual query text (a standard
+~4-characters-per-token heuristic) and the reduction percentage is
+computed from the selected `optimization_strategy` — not a fixed number
+returned regardless of input. See [`examples/basic_usage.py`](examples/basic_usage.py)
+for a full runnable walkthrough, including framework adapters for
+LangChain, LlamaIndex, CrewAI, Semantic Kernel, PydanticAI, and Haystack
+under [`pystreammcp.integrations`](python/pystreammcp/integrations/).
+
+## MCP orchestration & federation
+
+`Orchestrator` (`pystreammcp.Orchestrator`) discovers and routes across
+other MCP-enabled projects you configure in `pystreammcp.toml`:
+
+```toml
+[federation]
+endpoints = [
+    "http://localhost:8765/mcp",
+    "http://localhost:8766/mcp",
+]
+```
+
+```python
+from pystreammcp import Orchestrator
+
+orch = Orchestrator()  # reads ./pystreammcp.toml
+result = orch.discover_mcp_projects()
+# {"projects": [{"project_name": ..., "endpoint": ..., "status": "healthy"|"unavailable", ...}], "total": N}
+```
+
+This actually probes each configured endpoint's real `tools/list` MCP
+method — with nothing configured, or nothing reachable, it honestly
+reports zero/unavailable projects rather than a fixed fixture list.
+`Orchestrator` also does lexical tool-relevance ranking
+(`rank_tools_by_relevance`) and capability matching
+(`detect_compatible_projects`) against whatever it actually discovers.
+
+To expose these (plus tool-routing and webhook-event handling) as MCP
+tools, `PyStreamMCPHandler` (`pystreammcp._mcp_tools`) wires the MCP
+tool-call surface directly to a live `Orchestrator`/`EventRouter`
+instance — every handler call is backed by real, inspectable state, not
+a hardcoded response.
+
+**Honest limitations**: a few orchestration capabilities are accepted at
+the API level but not fully implemented yet — cross-project federated
+query *execution* and cross-database joins currently report
+`"status": "not_implemented"` with an explanation, rather than fabricated
+result rows, because there's no real query engine behind them yet.
+
+## Webhook event server
+
+`pystreammcp.server.create_flask_app()` exposes a small REST/webhook
+surface for orchestration events (agent MCPs reporting `mcp.available`,
+`tool.invoked`, health updates, etc. to `POST /orchestration/webhooks/events`).
+
+### Webhook security (HMAC-SHA256)
+
+Inbound events to `/orchestration/webhooks/events` must be signed:
+
+```bash
+export PYSTREAMMCP_WEBHOOK_SECRET="a long random shared secret"
+```
+
+Senders compute `HMAC-SHA256(secret, raw_request_body)` and send it as:
+
+```
+X-PyStreamMCP-Signature: sha256=<hex digest>
+```
+
+```python
+from pystreammcp.server import compute_webhook_signature
+import json, requests
+
+body = json.dumps({"event_type": "mcp.available", "data": {...}}).encode()
+signature = compute_webhook_signature(secret, body)
+
+requests.post(
+    "http://localhost:8000/orchestration/webhooks/events",
+    data=body,
+    headers={
+        "Content-Type": "application/json",
+        "X-PyStreamMCP-Signature": signature,
+    },
+)
+```
+
+Requests with a missing or invalid signature are rejected with `401`. If
+`PYSTREAMMCP_WEBHOOK_SECRET` isn't set at all, the endpoint fails closed
+and rejects every request with `503` — there is no "accept unsigned
+events" fallback.
+
+### Network defaults
+
+The HTTP/CLI servers (`pystreammcp server`, `PyStreamMCPServer`,
+`PyStreamMCPAPI.run`) bind to `127.0.0.1` (localhost-only) by default.
+Pass `--host 0.0.0.0` (CLI) or `host="0.0.0.0"` explicitly if you need the
+server reachable from other hosts — e.g. inside a container that already
+has its own network boundary.
+
+## Testing
+
+```bash
+pip install -e ".[dev,api,mcp,langchain,llamaindex,semantic-kernel]"
+pytest tests/ -v
+```
+
+## Rust workspace (not shipped)
+
+`core/` and `python/src/lib.rs` are an in-progress Rust workspace intended
+as a future performance backend. **It is not part of the published PyPI
+package** — the wheel built from `pyproject.toml` is pure Python — and it
+does not currently compile. Don't rely on it; it's tracked separately
+from the Python package described in this README.
+
+## License
+
+Proprietary License — Free to use with explicit attribution. See
+[LICENSE](LICENSE).

pystreammcp-3.2.0/README.md

@@ -0,0 +1,167 @@
+# PyStreamMCP
+
+[![CI](https://github.com/Mullassery/PyStreamMCP/actions/workflows/ci.yml/badge.svg)](https://github.com/Mullassery/PyStreamMCP/actions/workflows/ci.yml)
+
+An AI-agent intelligence layer: query planning, context discovery, and
+token-cost optimization, exposed both as a Python SDK and as an MCP
+(Model Context Protocol) tool server, plus a small event-driven
+orchestration/federation layer for coordinating multiple MCP endpoints.
+
+Pure Python — `pip install` just works, no Rust toolchain or compiler
+required.
+
+## Install
+
+```bash
+pip install PyStreamMCP
+```
+
+Optional extras, install what you actually need:
+
+```bash
+pip install "PyStreamMCP[api]"          # FastAPI + Flask HTTP servers
+pip install "PyStreamMCP[mcp]"          # MCP protocol client library
+pip install "PyStreamMCP[langchain]"    # LangChain adapter
+pip install "PyStreamMCP[llamaindex]"   # LlamaIndex adapter
+pip install "PyStreamMCP[semantic-kernel]"  # Semantic Kernel adapter
+pip install "PyStreamMCP[all-integrations]" # every framework adapter
+```
+
+Requires Python 3.9+.
+
+## Quick start
+
+```python
+from pystreammcp import Agent
+
+agent = Agent(
+    agent_id="recommendation_engine",
+    name="Product Recommendation Engine",
+    optimization_strategy="token_efficient",  # or "balanced" / "quality_first"
+    max_tokens=1000,
+)
+
+result = agent.query("What are the top 10 customers by lifetime value?")
+
+print(f"Baseline tokens:  {result.baseline_tokens}")
+print(f"Optimized tokens: {result.optimized_tokens}")
+print(f"Reduction:        {result.cost_reduction_percent:.1f}%")
+```
+
+Token counts are estimated from the actual query text (a standard
+~4-characters-per-token heuristic) and the reduction percentage is
+computed from the selected `optimization_strategy` — not a fixed number
+returned regardless of input. See [`examples/basic_usage.py`](examples/basic_usage.py)
+for a full runnable walkthrough, including framework adapters for
+LangChain, LlamaIndex, CrewAI, Semantic Kernel, PydanticAI, and Haystack
+under [`pystreammcp.integrations`](python/pystreammcp/integrations/).
+
+## MCP orchestration & federation
+
+`Orchestrator` (`pystreammcp.Orchestrator`) discovers and routes across
+other MCP-enabled projects you configure in `pystreammcp.toml`:
+
+```toml
+[federation]
+endpoints = [
+    "http://localhost:8765/mcp",
+    "http://localhost:8766/mcp",
+]
+```
+
+```python
+from pystreammcp import Orchestrator
+
+orch = Orchestrator()  # reads ./pystreammcp.toml
+result = orch.discover_mcp_projects()
+# {"projects": [{"project_name": ..., "endpoint": ..., "status": "healthy"|"unavailable", ...}], "total": N}
+```
+
+This actually probes each configured endpoint's real `tools/list` MCP
+method — with nothing configured, or nothing reachable, it honestly
+reports zero/unavailable projects rather than a fixed fixture list.
+`Orchestrator` also does lexical tool-relevance ranking
+(`rank_tools_by_relevance`) and capability matching
+(`detect_compatible_projects`) against whatever it actually discovers.
+
+To expose these (plus tool-routing and webhook-event handling) as MCP
+tools, `PyStreamMCPHandler` (`pystreammcp._mcp_tools`) wires the MCP
+tool-call surface directly to a live `Orchestrator`/`EventRouter`
+instance — every handler call is backed by real, inspectable state, not
+a hardcoded response.
+
+**Honest limitations**: a few orchestration capabilities are accepted at
+the API level but not fully implemented yet — cross-project federated
+query *execution* and cross-database joins currently report
+`"status": "not_implemented"` with an explanation, rather than fabricated
+result rows, because there's no real query engine behind them yet.
+
+## Webhook event server
+
+`pystreammcp.server.create_flask_app()` exposes a small REST/webhook
+surface for orchestration events (agent MCPs reporting `mcp.available`,
+`tool.invoked`, health updates, etc. to `POST /orchestration/webhooks/events`).
+
+### Webhook security (HMAC-SHA256)
+
+Inbound events to `/orchestration/webhooks/events` must be signed:
+
+```bash
+export PYSTREAMMCP_WEBHOOK_SECRET="a long random shared secret"
+```
+
+Senders compute `HMAC-SHA256(secret, raw_request_body)` and send it as:
+
+```
+X-PyStreamMCP-Signature: sha256=<hex digest>
+```
+
+```python
+from pystreammcp.server import compute_webhook_signature
+import json, requests
+
+body = json.dumps({"event_type": "mcp.available", "data": {...}}).encode()
+signature = compute_webhook_signature(secret, body)
+
+requests.post(
+    "http://localhost:8000/orchestration/webhooks/events",
+    data=body,
+    headers={
+        "Content-Type": "application/json",
+        "X-PyStreamMCP-Signature": signature,
+    },
+)
+```
+
+Requests with a missing or invalid signature are rejected with `401`. If
+`PYSTREAMMCP_WEBHOOK_SECRET` isn't set at all, the endpoint fails closed
+and rejects every request with `503` — there is no "accept unsigned
+events" fallback.
+
+### Network defaults
+
+The HTTP/CLI servers (`pystreammcp server`, `PyStreamMCPServer`,
+`PyStreamMCPAPI.run`) bind to `127.0.0.1` (localhost-only) by default.
+Pass `--host 0.0.0.0` (CLI) or `host="0.0.0.0"` explicitly if you need the
+server reachable from other hosts — e.g. inside a container that already
+has its own network boundary.
+
+## Testing
+
+```bash
+pip install -e ".[dev,api,mcp,langchain,llamaindex,semantic-kernel]"
+pytest tests/ -v
+```
+
+## Rust workspace (not shipped)
+
+`core/` and `python/src/lib.rs` are an in-progress Rust workspace intended
+as a future performance backend. **It is not part of the published PyPI
+package** — the wheel built from `pyproject.toml` is pure Python — and it
+does not currently compile. Don't rely on it; it's tracked separately
+from the Python package described in this README.
+
+## License
+
+Proprietary License — Free to use with explicit attribution. See
+[LICENSE](LICENSE).

{pystreammcp-2.1.0 → pystreammcp-3.2.0}/pyproject.toml

@@ -4,13 +4,20 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "PyStreamMCP"
-version = "2.1.0"
+version = "3.2.0"
 description = "Intelligent MCP orchestration hub - Intent understanding, capability matching, tool ranking with cohesive foundation"
 readme = "README.md"
-license = { text = "Proprietary" }
+license = {text = "Proprietary License — Free to use with explicit attribution"}
 authors = [{ name = "Georgi Mammen Mullassery", email = "mullassery@gmail.com" }]
 requires-python = ">=3.9"
-dependencies = ["pydantic>=2.0", "python-frontmatter>=1.0.0", "click>=8.0", "rich>=13.0"]
+dependencies = [
+    "pydantic>=2.0",
+    "python-frontmatter>=1.0.0",
+    "click>=8.0",
+    "rich>=13.0",
+    "httpx>=0.24.0",
+    "toml>=0.10; python_version < \"3.11\"",
+]
 classifiers = [
     "Development Status :: 5 - Production/Stable",
     "Intended Audience :: Developers",
@@ -39,6 +46,7 @@ api = [
     "fastapi>=0.104.0",
     "uvicorn[standard]>=0.24.0",
     "pydantic-settings>=2.0",
+    "flask>=2.3.0",
 ]
 
 # MCP (Model Context Protocol)

pystreammcp-3.2.0/python/PyStreamMCP.egg-info/PKG-INFO

@@ -0,0 +1,231 @@
+Metadata-Version: 2.4
+Name: PyStreamMCP
+Version: 3.2.0
+Summary: Intelligent MCP orchestration hub - Intent understanding, capability matching, tool ranking with cohesive foundation
+Author-email: Georgi Mammen Mullassery <mullassery@gmail.com>
+License: Proprietary License — Free to use with explicit attribution
+Project-URL: Repository, https://github.com/Mullassery/PyStreamMCP
+Project-URL: Issues, https://github.com/Mullassery/PyStreamMCP/issues
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: License :: Other/Proprietary License
+Classifier: Programming Language :: Python :: 3
+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: Programming Language :: Python :: 3.13
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: pydantic>=2.0
+Requires-Dist: python-frontmatter>=1.0.0
+Requires-Dist: click>=8.0
+Requires-Dist: rich>=13.0
+Requires-Dist: httpx>=0.24.0
+Requires-Dist: toml>=0.10; python_version < "3.11"
+Provides-Extra: dev
+Requires-Dist: pytest>=7.0; extra == "dev"
+Requires-Dist: pytest-cov>=4.0; extra == "dev"
+Requires-Dist: pytest-asyncio>=0.21.0; extra == "dev"
+Provides-Extra: api
+Requires-Dist: fastapi>=0.104.0; extra == "api"
+Requires-Dist: uvicorn[standard]>=0.24.0; extra == "api"
+Requires-Dist: pydantic-settings>=2.0; extra == "api"
+Requires-Dist: flask>=2.3.0; extra == "api"
+Provides-Extra: mcp
+Requires-Dist: mcp>=0.1.0; extra == "mcp"
+Provides-Extra: langchain
+Requires-Dist: langchain>=0.1.0; extra == "langchain"
+Provides-Extra: llamaindex
+Requires-Dist: llama-index>=0.9.0; extra == "llamaindex"
+Provides-Extra: semantic-kernel
+Requires-Dist: semantic-kernel>=0.4.0; extra == "semantic-kernel"
+Provides-Extra: crewai
+Requires-Dist: crewai>=0.1.0; extra == "crewai"
+Provides-Extra: pydantic-ai
+Requires-Dist: pydantic-ai>=0.1.0; extra == "pydantic-ai"
+Provides-Extra: haystack
+Requires-Dist: haystack-ai>=2.0.0; extra == "haystack"
+Provides-Extra: temporal
+Requires-Dist: temporalio>=1.0.0; extra == "temporal"
+Provides-Extra: airflow
+Requires-Dist: apache-airflow>=2.7.0; extra == "airflow"
+Provides-Extra: all-integrations
+Requires-Dist: langchain>=0.1.0; extra == "all-integrations"
+Requires-Dist: llama-index>=0.9.0; extra == "all-integrations"
+Requires-Dist: semantic-kernel>=0.4.0; extra == "all-integrations"
+Requires-Dist: crewai>=0.1.0; extra == "all-integrations"
+Requires-Dist: pydantic-ai>=0.1.0; extra == "all-integrations"
+Requires-Dist: haystack-ai>=2.0.0; extra == "all-integrations"
+Requires-Dist: temporalio>=1.0.0; extra == "all-integrations"
+Requires-Dist: apache-airflow>=2.7.0; extra == "all-integrations"
+Dynamic: license-file
+
+# PyStreamMCP
+
+[![CI](https://github.com/Mullassery/PyStreamMCP/actions/workflows/ci.yml/badge.svg)](https://github.com/Mullassery/PyStreamMCP/actions/workflows/ci.yml)
+
+An AI-agent intelligence layer: query planning, context discovery, and
+token-cost optimization, exposed both as a Python SDK and as an MCP
+(Model Context Protocol) tool server, plus a small event-driven
+orchestration/federation layer for coordinating multiple MCP endpoints.
+
+Pure Python — `pip install` just works, no Rust toolchain or compiler
+required.
+
+## Install
+
+```bash
+pip install PyStreamMCP
+```
+
+Optional extras, install what you actually need:
+
+```bash
+pip install "PyStreamMCP[api]"          # FastAPI + Flask HTTP servers
+pip install "PyStreamMCP[mcp]"          # MCP protocol client library
+pip install "PyStreamMCP[langchain]"    # LangChain adapter
+pip install "PyStreamMCP[llamaindex]"   # LlamaIndex adapter
+pip install "PyStreamMCP[semantic-kernel]"  # Semantic Kernel adapter
+pip install "PyStreamMCP[all-integrations]" # every framework adapter
+```
+
+Requires Python 3.9+.
+
+## Quick start
+
+```python
+from pystreammcp import Agent
+
+agent = Agent(
+    agent_id="recommendation_engine",
+    name="Product Recommendation Engine",
+    optimization_strategy="token_efficient",  # or "balanced" / "quality_first"
+    max_tokens=1000,
+)
+
+result = agent.query("What are the top 10 customers by lifetime value?")
+
+print(f"Baseline tokens:  {result.baseline_tokens}")
+print(f"Optimized tokens: {result.optimized_tokens}")
+print(f"Reduction:        {result.cost_reduction_percent:.1f}%")
+```
+
+Token counts are estimated from the actual query text (a standard
+~4-characters-per-token heuristic) and the reduction percentage is
+computed from the selected `optimization_strategy` — not a fixed number
+returned regardless of input. See [`examples/basic_usage.py`](examples/basic_usage.py)
+for a full runnable walkthrough, including framework adapters for
+LangChain, LlamaIndex, CrewAI, Semantic Kernel, PydanticAI, and Haystack
+under [`pystreammcp.integrations`](python/pystreammcp/integrations/).
+
+## MCP orchestration & federation
+
+`Orchestrator` (`pystreammcp.Orchestrator`) discovers and routes across
+other MCP-enabled projects you configure in `pystreammcp.toml`:
+
+```toml
+[federation]
+endpoints = [
+    "http://localhost:8765/mcp",
+    "http://localhost:8766/mcp",
+]
+```
+
+```python
+from pystreammcp import Orchestrator
+
+orch = Orchestrator()  # reads ./pystreammcp.toml
+result = orch.discover_mcp_projects()
+# {"projects": [{"project_name": ..., "endpoint": ..., "status": "healthy"|"unavailable", ...}], "total": N}
+```
+
+This actually probes each configured endpoint's real `tools/list` MCP
+method — with nothing configured, or nothing reachable, it honestly
+reports zero/unavailable projects rather than a fixed fixture list.
+`Orchestrator` also does lexical tool-relevance ranking
+(`rank_tools_by_relevance`) and capability matching
+(`detect_compatible_projects`) against whatever it actually discovers.
+
+To expose these (plus tool-routing and webhook-event handling) as MCP
+tools, `PyStreamMCPHandler` (`pystreammcp._mcp_tools`) wires the MCP
+tool-call surface directly to a live `Orchestrator`/`EventRouter`
+instance — every handler call is backed by real, inspectable state, not
+a hardcoded response.
+
+**Honest limitations**: a few orchestration capabilities are accepted at
+the API level but not fully implemented yet — cross-project federated
+query *execution* and cross-database joins currently report
+`"status": "not_implemented"` with an explanation, rather than fabricated
+result rows, because there's no real query engine behind them yet.
+
+## Webhook event server
+
+`pystreammcp.server.create_flask_app()` exposes a small REST/webhook
+surface for orchestration events (agent MCPs reporting `mcp.available`,
+`tool.invoked`, health updates, etc. to `POST /orchestration/webhooks/events`).
+
+### Webhook security (HMAC-SHA256)
+
+Inbound events to `/orchestration/webhooks/events` must be signed:
+
+```bash
+export PYSTREAMMCP_WEBHOOK_SECRET="a long random shared secret"
+```
+
+Senders compute `HMAC-SHA256(secret, raw_request_body)` and send it as:
+
+```
+X-PyStreamMCP-Signature: sha256=<hex digest>
+```
+
+```python
+from pystreammcp.server import compute_webhook_signature
+import json, requests
+
+body = json.dumps({"event_type": "mcp.available", "data": {...}}).encode()
+signature = compute_webhook_signature(secret, body)
+
+requests.post(
+    "http://localhost:8000/orchestration/webhooks/events",
+    data=body,
+    headers={
+        "Content-Type": "application/json",
+        "X-PyStreamMCP-Signature": signature,
+    },
+)
+```
+
+Requests with a missing or invalid signature are rejected with `401`. If
+`PYSTREAMMCP_WEBHOOK_SECRET` isn't set at all, the endpoint fails closed
+and rejects every request with `503` — there is no "accept unsigned
+events" fallback.
+
+### Network defaults
+
+The HTTP/CLI servers (`pystreammcp server`, `PyStreamMCPServer`,
+`PyStreamMCPAPI.run`) bind to `127.0.0.1` (localhost-only) by default.
+Pass `--host 0.0.0.0` (CLI) or `host="0.0.0.0"` explicitly if you need the
+server reachable from other hosts — e.g. inside a container that already
+has its own network boundary.
+
+## Testing
+
+```bash
+pip install -e ".[dev,api,mcp,langchain,llamaindex,semantic-kernel]"
+pytest tests/ -v
+```
+
+## Rust workspace (not shipped)
+
+`core/` and `python/src/lib.rs` are an in-progress Rust workspace intended
+as a future performance backend. **It is not part of the published PyPI
+package** — the wheel built from `pyproject.toml` is pure Python — and it
+does not currently compile. Don't rely on it; it's tracked separately
+from the Python package described in this README.
+
+## License
+
+Proprietary License — Free to use with explicit attribution. See
+[LICENSE](LICENSE).