PyPI - hexgate - Versions diffs - 0.2.2__tar.gz → 0.2.3__tar.gz - Mend

hexgate 0.2.2tar.gz → 0.2.3tar.gz

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.

Files changed (107) hide show

{hexgate-0.2.2 → hexgate-0.2.3}/PKG-INFO RENAMED Viewed

@@ -1,7 +1,7 @@
 Metadata-Version: 2.4
 Name: hexgate
-Version: 0.2.2
-Summary: HexaGate — authorization infrastructure for AI agents (agent runtime + cloud client).
+Version: 0.2.3
+Summary: Hexgate — authorization infrastructure for AI agents (agent runtime + cloud client).
 License-Expression: MIT
 Requires-Python: >=3.13
 Description-Content-Type: text/markdown
@@ -39,29 +39,38 @@ Dynamic: license-file
 <div align="center">
-# HexaGate
+<img src="./icon.svg" alt="Hexgate" width="96" height="96" />
+# Hexgate
 **Authorization infrastructure for AI agents.**
 Policy enforcement, signed policy bundles, per-request user scope, audit trail — for OpenAI Agents, LangChain, Google ADK, Pydantic AI, or a native runtime.
+[**Website**](https://hexgate.ai) · [Docs](https://docs.hexgate.ai) · [PyPI](https://pypi.org/project/hexgate/) · [Discussions](https://github.com/HexamindOrganisation/hexgate/discussions)
 [![PyPI](https://img.shields.io/pypi/v/hexgate?color=blue&logo=pypi&logoColor=white)](https://pypi.org/project/hexgate/)
-[![Python](https://img.shields.io/pypi/pyversions/hexgate?logo=python&logoColor=white)](https://pypi.org/project/hexgate/)
 [![CI](https://github.com/HexamindOrganisation/hexgate/actions/workflows/tests.yml/badge.svg?branch=main)](https://github.com/HexamindOrganisation/hexgate/actions/workflows/tests.yml)
 [![Downloads](https://img.shields.io/pypi/dm/hexgate?color=blueviolet)](https://pypi.org/project/hexgate/)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
-[Quick Start](#-quick-start--local-cli) · [Two paths](#-which-path-do-i-pick) · [Framework adapters](#-framework-agent-wrapping) · [Policy bundles](#-policy-bundles--compile-sign-enforce-wasm) · [User scope](#-user-scope--roles) · [Platform](#-hexagate-platform)
+<br />
+<img src="./assets/hero.png" alt="Control what your agents do — not just what they say. Policy decisions streaming live from the PolicyEnforcer." />
+<br />
+[Quick Start](#-quick-start--local-cli) · [Two paths](#-which-path-do-i-pick) · [Framework adapters](#-framework-agent-wrapping) · [Policy bundles](#-policy-bundles--compile-sign-enforce-wasm) · [User scope](#-user-scope--roles) · [Platform](#-hexgate-platform)
 </div>
 ---
-## What is HexaGate?
+## What is Hexgate?
-HexaGate is two things that move together:
+Hexgate is two things that move together:
 - **`hexgate` — the SDK.** A Python runtime that gates every tool call through a typed `Decision` (allow / deny / approval-required), wraps your existing OpenAI / LangChain / Google ADK / Pydantic AI agent without rewriting it, and threads per-request user identity through tracing + audit.
-- **The HexaGate platform** *(optional)* — a FastAPI control plane + React dashboard for editing policy in a browser, minting per-project tokens, watching live decisions stream from a serving agent, and shipping signed WASM policy bundles to production.
+- **The Hexgate platform** *(optional)* — a FastAPI control plane + React dashboard for editing policy in a browser, minting per-project tokens, watching live decisions stream from a serving agent, and shipping signed WASM policy bundles to production.
 You can use the SDK with nothing else (single-process REPL, YAML on disk). Or plug in the platform when you want auditable decisions in ClickHouse, a shared Playground UI, and live policy edits.
@@ -79,7 +88,7 @@ You can use the SDK with nothing else (single-process REPL, YAML on disk). Or pl
         ┌────────────────┐       ┌──────────────────┐       ┌────────────────┐
         │  Local policy  │       │ Signed WASM      │       │   Audit log    │
         │  (YAML / dir,  │       │ bundle from      │       │   (ClickHouse  │
-        │  hot reload)   │       │ HexaGate cloud   │       │   via REST)    │
+        │  hot reload)   │       │ Hexgate cloud   │       │   via REST)    │
         └────────────────┘       └──────────────────┘       └────────────────┘
 ```
@@ -103,7 +112,7 @@ You can use the SDK with nothing else (single-process REPL, YAML on disk). Or pl
 - [Environment](#-environment)
 - [Tests & dev tooling](#-tests--dev-tooling)
 - [CLI reference](#-cli-reference)
-- [HexaGate platform](#-hexagate-platform)
+- [Hexgate platform](#-hexgate-platform)
 - [User scope + roles](#-user-scope--roles)
 - [Stream results](#-stream-results)
@@ -177,7 +186,7 @@ Both commands accept either a plain agent id (`--agent researcher`) or a uvicorn
 ## 🚀 Quick Start — Platform
-To run the full HexaGate control plane locally (FastAPI backend + dashboard + your local agent serving over WebSocket), you need **three terminals**. The Makefile has a target that prints the recipe:
+To run the full Hexgate control plane locally (FastAPI backend + dashboard + your local agent serving over WebSocket), you need **three terminals**. The Makefile has a target that prints the recipe:
 ```bash
 make demo-platform     # prints the 3-terminal recipe below
@@ -660,7 +669,7 @@ What happens under the hood:
 Working scripts in `examples/`:
-- `examples/customer_bot.py` — canonical HexaGate path: `create_agent(...)` + the dashboard register/serve loop end-to-end.
+- `examples/customer_bot.py` — canonical Hexgate path: `create_agent(...)` + the dashboard register/serve loop end-to-end.
 - `examples/openai_demo.py` — `HexgateRunner` (OpenAI Agents SDK) end-to-end.
 - `examples/google_demo.py` — `HexgateRunner` (Google ADK) end-to-end with `InMemorySessionService`.
 - `examples/pydantic_ai_demo.py` — `wrap_pydantic_agent` (Pydantic AI) end-to-end.
@@ -795,7 +804,7 @@ That means the same agent code can stay simple in development, while deployment
 ## 🧩 Policy Bundles — Compile, Sign, Enforce (WASM)
-HexaGate has **two policy enforcement engines** that return identical decisions (there's a parity test suite that proves it):
+Hexgate has **two policy enforcement engines** that return identical decisions (there's a parity test suite that proves it):
 - **pydantic** (default) — evaluates constraints in-process. Zero setup; this is what every example above uses.
 - **WASM** — compiles `policy.yaml` → Rego → a WebAssembly module evaluated via `wasmtime`. This is the path production ships: one compiled artifact, byte-for-byte reproducible, cryptographically signed by the platform.
@@ -1131,7 +1140,7 @@ hexgate chat --use examples/research_agents.py --agent update_researcher --appro
 ### `hexgate register` — push a manifest to the platform
-Register a code-defined agent's manifest with the HexaGate platform. `--agent`
+Register a code-defined agent's manifest with the Hexgate platform. `--agent`
 takes a Python import path of the form `module.path:attribute`, the same shape
 as ASGI/WSGI entrypoints. The CLI imports the module, grabs the agent object,
 and POSTs its manifest to `${HEXGATE_API_URL}/v1/agents` using
@@ -1170,7 +1179,7 @@ system prompt directly off the object. No flags needed.
 `--system-prompt` accepts either a literal string or a path to a `.md` /
 `.txt` / `.jinja` file (read as text at register time).
-Supported frameworks: OpenAI Agents SDK, Google ADK, Pydantic AI, LangChain/LangGraph, HexaGate agents.
+Supported frameworks: OpenAI Agents SDK, Google ADK, Pydantic AI, LangChain/LangGraph, Hexgate agents.
 ### `hexgate serve` — bridge a local agent to the platform's relay
@@ -1207,7 +1216,7 @@ print(manifest.model_dump())
 ```
 `create_manifest` dispatches on the framework of `agent`. The supported
-types are the same set `hexgate register` accepts: HexaGate, OpenAI Agents
+types are the same set `hexgate register` accepts: Hexgate, OpenAI Agents
 SDK, Google ADK, Pydantic AI, and LangChain/LangGraph compiled graphs.
 For LangGraph you must pass `tools=` explicitly, and may pass `model=` /
 `system_prompt=`, since compiled graphs don't expose those fields after
@@ -1217,7 +1226,7 @@ The return value is an `AgentManifest` (a Pydantic model, also re-exported
 from `hexgate` for type annotations) — the same schema the platform
 stores and the dashboard renders.
-## 🌐 HexaGate Platform
+## 🌐 Hexgate Platform
 The `platform/` directory contains an optional control plane that hosts agent definitions, dev tokens, and a live debug surface. The SDK works fully without it (`load_local_agent`, `load_builtin_agent` keep their existing semantics) — but with it you get:
@@ -1336,7 +1345,7 @@ the name from the loaded agent's `.name` attribute — no env var needed.
 ## 👤 User Scope + Roles
-Real backends serve many users, and different users get different capabilities. HexaGate splits that into two pieces:
+Real backends serve many users, and different users get different capabilities. Hexgate splits that into two pieces:
 - **`User`** — the per-request scope. Marks "this invocation acts on behalf of alice, in role X." Async context manager; pushes a fact-bearing Biscuit through the agent runtime.
 - **Role policies** — one `policy.yaml` per role, optionally inheriting from a base mixin. The runtime picks the right one at call time based on the active `User.role`.
@@ -1486,3 +1495,7 @@ async for event in stream_agent(agent, handler, "latest AI breakthroughs"):
 - assistant text deltas
 - tool lifecycle
 - final run completion
+---
+If Hexgate looks useful, [give it a ⭐ on GitHub](https://github.com/HexamindOrganisation/hexgate) — it helps more than you'd think. Built by [Hexamind](https://hexgate.ai).

{hexgate-0.2.2 → hexgate-0.2.3}/README.md RENAMED Viewed

@@ -1,28 +1,37 @@
 <div align="center">
-# HexaGate
+<img src="./icon.svg" alt="Hexgate" width="96" height="96" />
+# Hexgate
 **Authorization infrastructure for AI agents.**
 Policy enforcement, signed policy bundles, per-request user scope, audit trail — for OpenAI Agents, LangChain, Google ADK, Pydantic AI, or a native runtime.
+[**Website**](https://hexgate.ai) · [Docs](https://docs.hexgate.ai) · [PyPI](https://pypi.org/project/hexgate/) · [Discussions](https://github.com/HexamindOrganisation/hexgate/discussions)
 [![PyPI](https://img.shields.io/pypi/v/hexgate?color=blue&logo=pypi&logoColor=white)](https://pypi.org/project/hexgate/)
-[![Python](https://img.shields.io/pypi/pyversions/hexgate?logo=python&logoColor=white)](https://pypi.org/project/hexgate/)
 [![CI](https://github.com/HexamindOrganisation/hexgate/actions/workflows/tests.yml/badge.svg?branch=main)](https://github.com/HexamindOrganisation/hexgate/actions/workflows/tests.yml)
 [![Downloads](https://img.shields.io/pypi/dm/hexgate?color=blueviolet)](https://pypi.org/project/hexgate/)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
-[Quick Start](#-quick-start--local-cli) · [Two paths](#-which-path-do-i-pick) · [Framework adapters](#-framework-agent-wrapping) · [Policy bundles](#-policy-bundles--compile-sign-enforce-wasm) · [User scope](#-user-scope--roles) · [Platform](#-hexagate-platform)
+<br />
+<img src="./assets/hero.png" alt="Control what your agents do — not just what they say. Policy decisions streaming live from the PolicyEnforcer." />
+<br />
+[Quick Start](#-quick-start--local-cli) · [Two paths](#-which-path-do-i-pick) · [Framework adapters](#-framework-agent-wrapping) · [Policy bundles](#-policy-bundles--compile-sign-enforce-wasm) · [User scope](#-user-scope--roles) · [Platform](#-hexgate-platform)
 </div>
 ---
-## What is HexaGate?
+## What is Hexgate?
-HexaGate is two things that move together:
+Hexgate is two things that move together:
 - **`hexgate` — the SDK.** A Python runtime that gates every tool call through a typed `Decision` (allow / deny / approval-required), wraps your existing OpenAI / LangChain / Google ADK / Pydantic AI agent without rewriting it, and threads per-request user identity through tracing + audit.
-- **The HexaGate platform** *(optional)* — a FastAPI control plane + React dashboard for editing policy in a browser, minting per-project tokens, watching live decisions stream from a serving agent, and shipping signed WASM policy bundles to production.
+- **The Hexgate platform** *(optional)* — a FastAPI control plane + React dashboard for editing policy in a browser, minting per-project tokens, watching live decisions stream from a serving agent, and shipping signed WASM policy bundles to production.
 You can use the SDK with nothing else (single-process REPL, YAML on disk). Or plug in the platform when you want auditable decisions in ClickHouse, a shared Playground UI, and live policy edits.
@@ -40,7 +49,7 @@ You can use the SDK with nothing else (single-process REPL, YAML on disk). Or pl
         ┌────────────────┐       ┌──────────────────┐       ┌────────────────┐
         │  Local policy  │       │ Signed WASM      │       │   Audit log    │
         │  (YAML / dir,  │       │ bundle from      │       │   (ClickHouse  │
-        │  hot reload)   │       │ HexaGate cloud   │       │   via REST)    │
+        │  hot reload)   │       │ Hexgate cloud   │       │   via REST)    │
         └────────────────┘       └──────────────────┘       └────────────────┘
 ```
@@ -64,7 +73,7 @@ You can use the SDK with nothing else (single-process REPL, YAML on disk). Or pl
 - [Environment](#-environment)
 - [Tests & dev tooling](#-tests--dev-tooling)
 - [CLI reference](#-cli-reference)
-- [HexaGate platform](#-hexagate-platform)
+- [Hexgate platform](#-hexgate-platform)
 - [User scope + roles](#-user-scope--roles)
 - [Stream results](#-stream-results)
@@ -138,7 +147,7 @@ Both commands accept either a plain agent id (`--agent researcher`) or a uvicorn
 ## 🚀 Quick Start — Platform
-To run the full HexaGate control plane locally (FastAPI backend + dashboard + your local agent serving over WebSocket), you need **three terminals**. The Makefile has a target that prints the recipe:
+To run the full Hexgate control plane locally (FastAPI backend + dashboard + your local agent serving over WebSocket), you need **three terminals**. The Makefile has a target that prints the recipe:
 ```bash
 make demo-platform     # prints the 3-terminal recipe below
@@ -621,7 +630,7 @@ What happens under the hood:
 Working scripts in `examples/`:
-- `examples/customer_bot.py` — canonical HexaGate path: `create_agent(...)` + the dashboard register/serve loop end-to-end.
+- `examples/customer_bot.py` — canonical Hexgate path: `create_agent(...)` + the dashboard register/serve loop end-to-end.
 - `examples/openai_demo.py` — `HexgateRunner` (OpenAI Agents SDK) end-to-end.
 - `examples/google_demo.py` — `HexgateRunner` (Google ADK) end-to-end with `InMemorySessionService`.
 - `examples/pydantic_ai_demo.py` — `wrap_pydantic_agent` (Pydantic AI) end-to-end.
@@ -756,7 +765,7 @@ That means the same agent code can stay simple in development, while deployment
 ## 🧩 Policy Bundles — Compile, Sign, Enforce (WASM)
-HexaGate has **two policy enforcement engines** that return identical decisions (there's a parity test suite that proves it):
+Hexgate has **two policy enforcement engines** that return identical decisions (there's a parity test suite that proves it):
 - **pydantic** (default) — evaluates constraints in-process. Zero setup; this is what every example above uses.
 - **WASM** — compiles `policy.yaml` → Rego → a WebAssembly module evaluated via `wasmtime`. This is the path production ships: one compiled artifact, byte-for-byte reproducible, cryptographically signed by the platform.
@@ -1092,7 +1101,7 @@ hexgate chat --use examples/research_agents.py --agent update_researcher --appro
 ### `hexgate register` — push a manifest to the platform
-Register a code-defined agent's manifest with the HexaGate platform. `--agent`
+Register a code-defined agent's manifest with the Hexgate platform. `--agent`
 takes a Python import path of the form `module.path:attribute`, the same shape
 as ASGI/WSGI entrypoints. The CLI imports the module, grabs the agent object,
 and POSTs its manifest to `${HEXGATE_API_URL}/v1/agents` using
@@ -1131,7 +1140,7 @@ system prompt directly off the object. No flags needed.
 `--system-prompt` accepts either a literal string or a path to a `.md` /
 `.txt` / `.jinja` file (read as text at register time).
-Supported frameworks: OpenAI Agents SDK, Google ADK, Pydantic AI, LangChain/LangGraph, HexaGate agents.
+Supported frameworks: OpenAI Agents SDK, Google ADK, Pydantic AI, LangChain/LangGraph, Hexgate agents.
 ### `hexgate serve` — bridge a local agent to the platform's relay
@@ -1168,7 +1177,7 @@ print(manifest.model_dump())
 ```
 `create_manifest` dispatches on the framework of `agent`. The supported
-types are the same set `hexgate register` accepts: HexaGate, OpenAI Agents
+types are the same set `hexgate register` accepts: Hexgate, OpenAI Agents
 SDK, Google ADK, Pydantic AI, and LangChain/LangGraph compiled graphs.
 For LangGraph you must pass `tools=` explicitly, and may pass `model=` /
 `system_prompt=`, since compiled graphs don't expose those fields after
@@ -1178,7 +1187,7 @@ The return value is an `AgentManifest` (a Pydantic model, also re-exported
 from `hexgate` for type annotations) — the same schema the platform
 stores and the dashboard renders.
-## 🌐 HexaGate Platform
+## 🌐 Hexgate Platform
 The `platform/` directory contains an optional control plane that hosts agent definitions, dev tokens, and a live debug surface. The SDK works fully without it (`load_local_agent`, `load_builtin_agent` keep their existing semantics) — but with it you get:
@@ -1297,7 +1306,7 @@ the name from the loaded agent's `.name` attribute — no env var needed.
 ## 👤 User Scope + Roles
-Real backends serve many users, and different users get different capabilities. HexaGate splits that into two pieces:
+Real backends serve many users, and different users get different capabilities. Hexgate splits that into two pieces:
 - **`User`** — the per-request scope. Marks "this invocation acts on behalf of alice, in role X." Async context manager; pushes a fact-bearing Biscuit through the agent runtime.
 - **Role policies** — one `policy.yaml` per role, optionally inheriting from a base mixin. The runtime picks the right one at call time based on the active `User.role`.
@@ -1447,3 +1456,7 @@ async for event in stream_agent(agent, handler, "latest AI breakthroughs"):
 - assistant text deltas
 - tool lifecycle
 - final run completion
+---
+If Hexgate looks useful, [give it a ⭐ on GitHub](https://github.com/HexamindOrganisation/hexgate) — it helps more than you'd think. Built by [Hexamind](https://hexgate.ai).

{hexgate-0.2.2 → hexgate-0.2.3}/hexgate/adapters/google/runner.py RENAMED Viewed

@@ -21,7 +21,7 @@ from hexgate.runtime import User
 class HexgateRunner:
-    """Runner for Google ADK agents with HexaGate tool policy and observability."""
+    """Runner for Google ADK agents with Hexgate tool policy and observability."""
     def __init__(
         self,

{hexgate-0.2.2 → hexgate-0.2.3}/hexgate/adapters/langchain/agent.py RENAMED Viewed

@@ -1,4 +1,4 @@
-"""Proxy around a pre-built ``CompiledStateGraph`` for HexaGate-aware calls."""
+"""Proxy around a pre-built ``CompiledStateGraph`` for Hexgate-aware calls."""
 from __future__ import annotations
@@ -60,7 +60,7 @@ class HexgateLangchainAgent:
         }
     def _with_callbacks(self, config: RunnableConfig | None) -> RunnableConfig:
-        """Append the HexaGate callback handler to ``config['callbacks']``."""
+        """Append the Hexgate callback handler to ``config['callbacks']``."""
         merged: RunnableConfig = dict(config) if config else {}
         callbacks = list(merged.get("callbacks") or [])
         if self._callback_handler not in callbacks:

{hexgate-0.2.2 → hexgate-0.2.3}/hexgate/adapters/langchain/wrapper.py RENAMED Viewed

@@ -1,5 +1,5 @@
 """BYO-graph entry point: retrofit a pre-built ``CompiledStateGraph`` with
-HexaGate policy. Tools are mutated in place so the graph keeps its
+Hexgate policy. Tools are mutated in place so the graph keeps its
 references; the returned :class:`HexgateLangchainAgent` opens a User
 scope + Langfuse propagation per call. For the manifest-driven path,
 use :func:`hexgate.enforce_policy` instead.
@@ -28,7 +28,7 @@ def wrap_langchain_agent(
     tools: list[BaseTool],
     api_key: str | None = None,
 ) -> HexgateLangchainAgent:
-    """Wrap a pre-built LangGraph agent with HexaGate policy enforcement.
+    """Wrap a pre-built LangGraph agent with Hexgate policy enforcement.
     Mutates ``tools`` in place so the graph keeps its references.
     The returned proxy takes ``user`` per invocation; role resolves at

{hexgate-0.2.2 → hexgate-0.2.3}/hexgate/adapters/openai/runner.py RENAMED Viewed

@@ -32,7 +32,7 @@ from hexgate.security.enforcer import build_enforcer
 class HexgateRunner:
-    """Runner for OpenAI agents with HexaGate tool policy and observability."""
+    """Runner for OpenAI agents with Hexgate tool policy and observability."""
     def __init__(self, api_key: str | None = None):
         self.api_key = api_key or os.getenv("HEXGATE_KEY")

{hexgate-0.2.2 → hexgate-0.2.3}/hexgate/adapters/pydantic_ai/agent.py RENAMED Viewed

@@ -1,4 +1,4 @@
-"""Proxy around a pydantic_ai ``Agent`` for HexaGate-aware calls."""
+"""Proxy around a pydantic_ai ``Agent`` for Hexgate-aware calls."""
 from __future__ import annotations

{hexgate-0.2.2 → hexgate-0.2.3}/hexgate/adapters/pydantic_ai/wrapper.py RENAMED Viewed

@@ -49,7 +49,7 @@ def wrap_pydantic_agent(
     agent: Agent,
     api_key: str | None = None,
 ) -> HexgatePydanticAgent:
-    """Wrap a pydantic_ai agent with HexaGate policy + observability.
+    """Wrap a pydantic_ai agent with Hexgate policy + observability.
     Returns a :class:`HexgatePydanticAgent` backed by a clone of the
     caller's ``agent``; the original is not mutated. The proxy takes

{hexgate-0.2.2 → hexgate-0.2.3}/hexgate/agents/loader.py RENAMED Viewed

@@ -462,10 +462,10 @@ def load_hexgate_agent(
     approval_handler: ApprovalHandler | None = None,
     decision_observer: "DecisionObserver | None" = None,
 ) -> tuple[AgentGraph, CallbackHandler]:
-    """Fetch an agent from HexaGate and return it with policy enforcement applied.
+    """Fetch an agent from Hexgate and return it with policy enforcement applied.
     Mirrors `load_local_agent` but sources the three YAMLs (agent, policy, system)
-    from the HexaGate API instead of disk. Tool resolution and enforcement are
+    from the Hexgate API instead of disk. Tool resolution and enforcement are
     identical — only the bytes' origin differs.
     ``name`` is required. The Phase-7 env-var fallback chain
@@ -555,7 +555,7 @@ def load_agent(
     approval_handler: ApprovalHandler | None = None,
     decision_observer: "DecisionObserver | None" = None,
 ) -> tuple[AgentGraph, CallbackHandler]:
-    """Load an agent from HexaGate (when HEXGATE_KEY is set), local, or builtin.
+    """Load an agent from Hexgate (when HEXGATE_KEY is set), local, or builtin.
     ``name`` is required for every path post-Phase 7 — the
     HEXGATE_AGENT_NAME env-var fallback was removed when ``hexgate
@@ -584,7 +584,7 @@ def load_agent(
             decision_observer=decision_observer,
         )
     if name is None:
-        raise ValueError("load_agent() requires a name when not using HexaGate Cloud")
+        raise ValueError("load_agent() requires a name when not using Hexgate Cloud")
     source = resolve_agent_source(name, base_dir)
     if source == "local":
         return load_local_agent(

{hexgate-0.2.2 → hexgate-0.2.3}/hexgate/cli/_common.py RENAMED Viewed

@@ -63,7 +63,7 @@ def build_runtime(
       "this loads in serve but not chat" footgun — same spec form
       ``hexgate serve`` already accepts.
-    ``local_only=True`` keeps the loader off the HexaGate Cloud path even
+    ``local_only=True`` keeps the loader off the Hexgate Cloud path even
     when ``HEXGATE_KEY`` is present in the environment — what terminal
     chat uses, since it doesn't need cloud-fetched policy or a serve
     tunnel. ``hexgate serve`` passes ``local_only=False`` so policy edits

{hexgate-0.2.2 → hexgate-0.2.3}/hexgate/cli/register/hexgate.py RENAMED Viewed

@@ -19,10 +19,10 @@ from hexgate.cli.register.models import (
 def create_hexgate_manifest(
     agent: HexgateAgent, *, description: str | None = None
 ) -> AgentManifest:
-    """Build an AgentManifest from a HexaGate agent created by `create_agent`."""
+    """Build an AgentManifest from a Hexgate agent created by `create_agent`."""
     if not agent.name:
         raise ValueError(
-            "HexaGate agent has no name — set a name on the HexgateAgent so the "
+            "Hexgate agent has no name — set a name on the HexgateAgent so the "
             "manifest can identify it on the platform."
         )
     return AgentManifest(

{hexgate-0.2.2 → hexgate-0.2.3}/hexgate/cli/register/main.py RENAMED Viewed

@@ -40,8 +40,8 @@ def add_parser(subparsers: argparse._SubParsersAction) -> None:
     """
     parser = subparsers.add_parser(
         "register",
-        help="Register an agent to the HexaGate platform.",
-        description="Register an agent to the HexaGate platform.",
+        help="Register an agent to the Hexgate platform.",
+        description="Register an agent to the Hexgate platform.",
     )
     parser.add_argument(
         "--agent",

{hexgate-0.2.2 → hexgate-0.2.3}/hexgate/cli/serve.py RENAMED Viewed

@@ -1,4 +1,4 @@
-"""Serve subcommand: bridge a local agent to the HexaGate control plane.
+"""Serve subcommand: bridge a local agent to the Hexgate control plane.
 Connects to ``ws://{API_URL}/v1/serve`` and authenticates via the
 ``bearer.<envelope>`` WebSocket subprotocol — the server derives the
@@ -251,9 +251,9 @@ def add_parser(subparsers: argparse._SubParsersAction) -> None:
     """Register the `serve` subcommand on the top-level hexgate CLI."""
     parser = subparsers.add_parser(
         "serve",
-        help="Relay a local agent to the HexaGate dashboard over WebSocket.",
+        help="Relay a local agent to the Hexgate dashboard over WebSocket.",
         description=(
-            "Serve a local agent to the HexaGate dashboard Playground over "
+            "Serve a local agent to the Hexgate dashboard Playground over "
             "WebSocket. Takes a module:attr spec — the same form as "
             "`hexgate register --agent ...` — and brings the agent up "
             "end-to-end: auto-registers the manifest (idempotent), fetches "

{hexgate-0.2.2 → hexgate-0.2.3}/hexgate/cloud/__init__.py RENAMED Viewed

@@ -1,4 +1,4 @@
-"""Transport layer for the HexaGate control plane — HTTP client + Biscuit verify.
+"""Transport layer for the Hexgate control plane — HTTP client + Biscuit verify.
 The agent-loader equivalent (`load_hexgate_agent`) lives in `hexgate.agents.loader`
 alongside `load_local_agent` / `load_builtin_agent`; this package only carries the

{hexgate-0.2.2 → hexgate-0.2.3}/hexgate/cloud/attenuate.py RENAMED Viewed

@@ -1,4 +1,4 @@
-"""Attenuate a verified HexaGate token with user-scoped facts and checks.
+"""Attenuate a verified Hexgate token with user-scoped facts and checks.
 In production the dev's backend calls :func:`attenuate_for_user` on each
 inbound request, taking the project-wide token from ``HEXGATE_KEY`` and

{hexgate-0.2.2 → hexgate-0.2.3}/hexgate/cloud/biscuit.py RENAMED Viewed

@@ -71,7 +71,7 @@ def parse_envelope(envelope: str) -> tuple[str, str, str]:
     parts = envelope.split("_", 3)
     if len(parts) != 4 or parts[0] != ENVELOPE_PREFIX:
         raise TokenError(
-            f"malformed HexaGate token envelope (expected '{ENVELOPE_PREFIX}_<env>_<project>_<biscuit>')"
+            f"malformed Hexgate token envelope (expected '{ENVELOPE_PREFIX}_<env>_<project>_<biscuit>')"
         )
     env, project_id, biscuit_b64 = parts[1], parts[2], parts[3]
     return env, project_id, biscuit_b64

{hexgate-0.2.2 → hexgate-0.2.3}/hexgate/cloud/client.py RENAMED Viewed

@@ -1,4 +1,4 @@
-"""HTTP client for the HexaGate control plane.
+"""HTTP client for the Hexgate control plane.
 The client trusts ``HEXGATE_KEY`` only after verifying its Biscuit signature
 against the platform's public key. The public key is resolved in this order:
@@ -6,7 +6,7 @@ against the platform's public key. The public key is resolved in this order:
 1. Explicit ``public_key`` arg passed to ``HexgateConfig``.
 2. ``HEXGATE_PUBLIC_KEY`` env var (base64 url-safe, 32 raw bytes).
 3. Fetched from ``GET /v1/.well-known/keys`` on first use (TOFU for POC;
-   embed a build-time constant for hosted HexaGate Cloud later).
+   embed a build-time constant for hosted Hexgate Cloud later).
 If none of the above produces a verifying key, the client raises with a
 clear error rather than blindly forwarding the bearer token.
@@ -42,7 +42,7 @@ TOKEN_PREFIX = "fty_"
 class HexgateError(RuntimeError):
-    """Raised for any HexaGate API interaction failure.
+    """Raised for any Hexgate API interaction failure.
     ``status`` is the HTTP status code, or ``None`` for transport errors.
     """
@@ -54,7 +54,7 @@ class HexgateError(RuntimeError):
 @dataclass
 class HexgateConfig:
-    """Resolved configuration for a HexaGate client.
+    """Resolved configuration for a Hexgate client.
     ``project_id`` is best-effort only — used by display surfaces
     (log lines, langchain tags) but never threaded through API URLs.
@@ -332,14 +332,14 @@ class HexgateClient:
                 return None, exc.headers.get("ETag") or if_none_match
             detail = exc.read().decode("utf-8", errors="replace")
             raise HexgateError(
-                f"HexaGate API error {exc.code} calling {url}: {detail[:200]}",
+                f"Hexgate API error {exc.code} calling {url}: {detail[:200]}",
                 status=exc.code,
             ) from exc
         except urllib.error.URLError as exc:
             raise HexgateError(
-                f"HexaGate API unreachable at {url}: {exc.reason}"
+                f"Hexgate API unreachable at {url}: {exc.reason}"
             ) from exc
         try:
             return json.loads(payload), etag
         except json.JSONDecodeError as exc:
-            raise HexgateError(f"HexaGate API returned non-JSON from {url}") from exc
+            raise HexgateError(f"Hexgate API returned non-JSON from {url}") from exc

{hexgate-0.2.2 → hexgate-0.2.3}/hexgate/security/rego.py RENAMED Viewed

@@ -1,4 +1,4 @@
-"""Compile a HexaGate ``policy.yaml`` document into a Rego policy module.
+"""Compile a Hexgate ``policy.yaml`` document into a Rego policy module.
 The compiler is a pure-Python transformation: parsed YAML payload →
 ``PolicySet`` (which flattens ``inherits:`` and drops ``is_mixin:`` entries)

{hexgate-0.2.2 → hexgate-0.2.3}/hexgate/security/wasm_engine.py RENAMED Viewed

@@ -1,4 +1,4 @@
-"""Evaluate a HexaGate policy bundle's ``policy.wasm`` via wasmtime-py.
+"""Evaluate a Hexgate policy bundle's ``policy.wasm`` via wasmtime-py.
 This module is the runtime counterpart to :mod:`hexgate.security.rego_wasm`:
 that one *compiles* YAML → Rego → WASM, this one *evaluates* WASM →
@@ -107,7 +107,7 @@ DEFAULT_ENTRYPOINT = "hexgate/policy/decision"
 class WasmPolicy:
-    """A loaded, evaluable HexaGate policy WASM bundle.
+    """A loaded, evaluable Hexgate policy WASM bundle.
     Construct with ``WasmPolicy.from_bytes(wasm)`` or
     ``WasmPolicy.from_bundle_path(dir)``. Each instance owns its own
@@ -349,7 +349,7 @@ def _make_builtin_stubs(store: wasmtime.Store) -> dict[int, wasmtime.Func]:
             builtin_id = args[0] if args else -1
             raise WasmEvalError(
                 f"policy invoked unsupported Rego builtin "
-                f"(builtin{_arity}, id={builtin_id}); the HexaGate "
+                f"(builtin{_arity}, id={builtin_id}); the Hexgate "
                 "constraint grammar does not expose builtins."
             )

{hexgate-0.2.2 → hexgate-0.2.3}/hexgate/streaming/__init__.py RENAMED Viewed

@@ -1,4 +1,4 @@
-"""Streaming events and the LangChain → HexaGate event normalizer."""
+"""Streaming events and the LangChain → Hexgate event normalizer."""
 from hexgate.streaming.events import (
     AgentRunResult,

{hexgate-0.2.2 → hexgate-0.2.3}/hexgate.egg-info/PKG-INFO RENAMED Viewed

@@ -1,7 +1,7 @@
 Metadata-Version: 2.4
 Name: hexgate
-Version: 0.2.2
-Summary: HexaGate — authorization infrastructure for AI agents (agent runtime + cloud client).
+Version: 0.2.3
+Summary: Hexgate — authorization infrastructure for AI agents (agent runtime + cloud client).
 License-Expression: MIT
 Requires-Python: >=3.13
 Description-Content-Type: text/markdown
@@ -39,29 +39,38 @@ Dynamic: license-file
 <div align="center">
-# HexaGate
+<img src="./icon.svg" alt="Hexgate" width="96" height="96" />
+# Hexgate
 **Authorization infrastructure for AI agents.**
 Policy enforcement, signed policy bundles, per-request user scope, audit trail — for OpenAI Agents, LangChain, Google ADK, Pydantic AI, or a native runtime.
+[**Website**](https://hexgate.ai) · [Docs](https://docs.hexgate.ai) · [PyPI](https://pypi.org/project/hexgate/) · [Discussions](https://github.com/HexamindOrganisation/hexgate/discussions)
 [![PyPI](https://img.shields.io/pypi/v/hexgate?color=blue&logo=pypi&logoColor=white)](https://pypi.org/project/hexgate/)
-[![Python](https://img.shields.io/pypi/pyversions/hexgate?logo=python&logoColor=white)](https://pypi.org/project/hexgate/)
 [![CI](https://github.com/HexamindOrganisation/hexgate/actions/workflows/tests.yml/badge.svg?branch=main)](https://github.com/HexamindOrganisation/hexgate/actions/workflows/tests.yml)
 [![Downloads](https://img.shields.io/pypi/dm/hexgate?color=blueviolet)](https://pypi.org/project/hexgate/)
 [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
-[Quick Start](#-quick-start--local-cli) · [Two paths](#-which-path-do-i-pick) · [Framework adapters](#-framework-agent-wrapping) · [Policy bundles](#-policy-bundles--compile-sign-enforce-wasm) · [User scope](#-user-scope--roles) · [Platform](#-hexagate-platform)
+<br />
+<img src="./assets/hero.png" alt="Control what your agents do — not just what they say. Policy decisions streaming live from the PolicyEnforcer." />
+<br />
+[Quick Start](#-quick-start--local-cli) · [Two paths](#-which-path-do-i-pick) · [Framework adapters](#-framework-agent-wrapping) · [Policy bundles](#-policy-bundles--compile-sign-enforce-wasm) · [User scope](#-user-scope--roles) · [Platform](#-hexgate-platform)
 </div>
 ---
-## What is HexaGate?
+## What is Hexgate?
-HexaGate is two things that move together:
+Hexgate is two things that move together:
 - **`hexgate` — the SDK.** A Python runtime that gates every tool call through a typed `Decision` (allow / deny / approval-required), wraps your existing OpenAI / LangChain / Google ADK / Pydantic AI agent without rewriting it, and threads per-request user identity through tracing + audit.
-- **The HexaGate platform** *(optional)* — a FastAPI control plane + React dashboard for editing policy in a browser, minting per-project tokens, watching live decisions stream from a serving agent, and shipping signed WASM policy bundles to production.
+- **The Hexgate platform** *(optional)* — a FastAPI control plane + React dashboard for editing policy in a browser, minting per-project tokens, watching live decisions stream from a serving agent, and shipping signed WASM policy bundles to production.
 You can use the SDK with nothing else (single-process REPL, YAML on disk). Or plug in the platform when you want auditable decisions in ClickHouse, a shared Playground UI, and live policy edits.
@@ -79,7 +88,7 @@ You can use the SDK with nothing else (single-process REPL, YAML on disk). Or pl
         ┌────────────────┐       ┌──────────────────┐       ┌────────────────┐
         │  Local policy  │       │ Signed WASM      │       │   Audit log    │
         │  (YAML / dir,  │       │ bundle from      │       │   (ClickHouse  │
-        │  hot reload)   │       │ HexaGate cloud   │       │   via REST)    │
+        │  hot reload)   │       │ Hexgate cloud   │       │   via REST)    │
         └────────────────┘       └──────────────────┘       └────────────────┘
 ```
@@ -103,7 +112,7 @@ You can use the SDK with nothing else (single-process REPL, YAML on disk). Or pl
 - [Environment](#-environment)
 - [Tests & dev tooling](#-tests--dev-tooling)
 - [CLI reference](#-cli-reference)
-- [HexaGate platform](#-hexagate-platform)
+- [Hexgate platform](#-hexgate-platform)
 - [User scope + roles](#-user-scope--roles)
 - [Stream results](#-stream-results)
@@ -177,7 +186,7 @@ Both commands accept either a plain agent id (`--agent researcher`) or a uvicorn
 ## 🚀 Quick Start — Platform
-To run the full HexaGate control plane locally (FastAPI backend + dashboard + your local agent serving over WebSocket), you need **three terminals**. The Makefile has a target that prints the recipe:
+To run the full Hexgate control plane locally (FastAPI backend + dashboard + your local agent serving over WebSocket), you need **three terminals**. The Makefile has a target that prints the recipe:
 ```bash
 make demo-platform     # prints the 3-terminal recipe below
@@ -660,7 +669,7 @@ What happens under the hood:
 Working scripts in `examples/`:
-- `examples/customer_bot.py` — canonical HexaGate path: `create_agent(...)` + the dashboard register/serve loop end-to-end.
+- `examples/customer_bot.py` — canonical Hexgate path: `create_agent(...)` + the dashboard register/serve loop end-to-end.
 - `examples/openai_demo.py` — `HexgateRunner` (OpenAI Agents SDK) end-to-end.
 - `examples/google_demo.py` — `HexgateRunner` (Google ADK) end-to-end with `InMemorySessionService`.
 - `examples/pydantic_ai_demo.py` — `wrap_pydantic_agent` (Pydantic AI) end-to-end.
@@ -795,7 +804,7 @@ That means the same agent code can stay simple in development, while deployment
 ## 🧩 Policy Bundles — Compile, Sign, Enforce (WASM)
-HexaGate has **two policy enforcement engines** that return identical decisions (there's a parity test suite that proves it):
+Hexgate has **two policy enforcement engines** that return identical decisions (there's a parity test suite that proves it):
 - **pydantic** (default) — evaluates constraints in-process. Zero setup; this is what every example above uses.
 - **WASM** — compiles `policy.yaml` → Rego → a WebAssembly module evaluated via `wasmtime`. This is the path production ships: one compiled artifact, byte-for-byte reproducible, cryptographically signed by the platform.
@@ -1131,7 +1140,7 @@ hexgate chat --use examples/research_agents.py --agent update_researcher --appro
 ### `hexgate register` — push a manifest to the platform
-Register a code-defined agent's manifest with the HexaGate platform. `--agent`
+Register a code-defined agent's manifest with the Hexgate platform. `--agent`
 takes a Python import path of the form `module.path:attribute`, the same shape
 as ASGI/WSGI entrypoints. The CLI imports the module, grabs the agent object,
 and POSTs its manifest to `${HEXGATE_API_URL}/v1/agents` using
@@ -1170,7 +1179,7 @@ system prompt directly off the object. No flags needed.
 `--system-prompt` accepts either a literal string or a path to a `.md` /
 `.txt` / `.jinja` file (read as text at register time).
-Supported frameworks: OpenAI Agents SDK, Google ADK, Pydantic AI, LangChain/LangGraph, HexaGate agents.
+Supported frameworks: OpenAI Agents SDK, Google ADK, Pydantic AI, LangChain/LangGraph, Hexgate agents.
 ### `hexgate serve` — bridge a local agent to the platform's relay
@@ -1207,7 +1216,7 @@ print(manifest.model_dump())
 ```
 `create_manifest` dispatches on the framework of `agent`. The supported
-types are the same set `hexgate register` accepts: HexaGate, OpenAI Agents
+types are the same set `hexgate register` accepts: Hexgate, OpenAI Agents
 SDK, Google ADK, Pydantic AI, and LangChain/LangGraph compiled graphs.
 For LangGraph you must pass `tools=` explicitly, and may pass `model=` /
 `system_prompt=`, since compiled graphs don't expose those fields after
@@ -1217,7 +1226,7 @@ The return value is an `AgentManifest` (a Pydantic model, also re-exported
 from `hexgate` for type annotations) — the same schema the platform
 stores and the dashboard renders.
-## 🌐 HexaGate Platform
+## 🌐 Hexgate Platform
 The `platform/` directory contains an optional control plane that hosts agent definitions, dev tokens, and a live debug surface. The SDK works fully without it (`load_local_agent`, `load_builtin_agent` keep their existing semantics) — but with it you get:
@@ -1336,7 +1345,7 @@ the name from the loaded agent's `.name` attribute — no env var needed.
 ## 👤 User Scope + Roles
-Real backends serve many users, and different users get different capabilities. HexaGate splits that into two pieces:
+Real backends serve many users, and different users get different capabilities. Hexgate splits that into two pieces:
 - **`User`** — the per-request scope. Marks "this invocation acts on behalf of alice, in role X." Async context manager; pushes a fact-bearing Biscuit through the agent runtime.
 - **Role policies** — one `policy.yaml` per role, optionally inheriting from a base mixin. The runtime picks the right one at call time based on the active `User.role`.
@@ -1486,3 +1495,7 @@ async for event in stream_agent(agent, handler, "latest AI breakthroughs"):
 - assistant text deltas
 - tool lifecycle
 - final run completion
+---
+If Hexgate looks useful, [give it a ⭐ on GitHub](https://github.com/HexamindOrganisation/hexgate) — it helps more than you'd think. Built by [Hexamind](https://hexgate.ai).

{hexgate-0.2.2 → hexgate-0.2.3}/pyproject.toml RENAMED Viewed

@@ -9,8 +9,8 @@ build-backend = "setuptools.build_meta"
 # 0.2.0 (the original package name was taken on PyPI by a 2014
 # abandoned project; the team consolidated on `hexgate` for everything).
 name = "hexgate"
-version = "0.2.2"
-description = "HexaGate — authorization infrastructure for AI agents (agent runtime + cloud client)."
+version = "0.2.3"
+description = "Hexgate — authorization infrastructure for AI agents (agent runtime + cloud client)."
 readme = "README.md"
 license = "MIT"
 license-files = ["LICENSE"]