Release 0.2.15

Author: lpalbou

CHANGELOG.md

@@ -7,6 +7,22 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.2.15] - 2026-05-21
+
+### Added
+- Added Runtime-backed durable bloc prompt-cache control-plane routes under `/api/gateway/blocs/*`, including KV manifest/list/ensure/load/delete/prune helpers for exact-reuse workflows.
+- Added Gateway-owned workspace file helper support plus focused route and contract coverage for durable blocs, model residency, notifier behavior, and Runtime-backed capability discovery.
+
+### Changed
+- Raised the Runtime floor to `AbstractRuntime[multimodal,mcp-worker]>=0.4.19` and moved Gateway's public provider/media/tool boundary behind Runtime facades rather than direct package imports.
+- Updated Apple/GPU install profiles to cascade through Runtime's aggregate extras and excluded internal `tests/`, `flows/`, and backlog notes from source distributions.
+- Expanded the docs and capability contract to cover durable blocs, media/model residency, Runtime-backed email/Telegram helpers, and the current Docker/runtime dependency shape.
+
+### Fixed
+- Gateway no longer reads AbstractCore config for LLM helper defaults; provider/model resolution now follows request values, Gateway env, and flow defaults with a clear config error when unset.
+- Gateway's operator email, Telegram, and notification paths now use Runtime's AbstractCore host facades, while local file/workspace helpers stay owned by Gateway.
+- Capability discovery and prompt-cache readiness reporting now better reflect the actual state of generated-media, voice/audio, and provider-backed cache controls.
+
 ## [0.2.14] - 2026-05-19
 
 ### Fixed

README.md

@@ -15,7 +15,7 @@ Start here: [docs/getting-started.md](docs/getting-started.md)
 AbstractGateway is part of the **AbstractFramework** ecosystem:
 
 - **AbstractRuntime** (required): durable run model + workflow registry + stores (`pyproject.toml`, `src/abstractgateway/runner.py`)
-- **AbstractCore / AbstractVoice / AbstractVision / AbstractMemory** (required by the default server install): LLM/tool execution, provider-level prompt-cache controls, workflow-backed/direct generated image/voice/audio capabilities, and KG memory used by Gateway bundles (`src/abstractgateway/hosts/bundle_host.py`)
+- **AbstractRuntime + transitive capability packages** (required by the default server install): Runtime owns the LLM/tool/media integration boundary; Gateway uses its discovery/run facades for prompt-cache controls, generated image/voice/audio capabilities, and KG-backed bundle execution (`src/abstractgateway/hosts/bundle_host.py`)
 - Higher-level UIs (optional): AbstractFlow (authoring/bundling), AbstractCode / AbstractObserver / thin clients (rendering + operations)
 
 Related repos:
@@ -56,24 +56,23 @@ Release images are published to GHCR. The default image is the light,
 portable server image:
 
 ```bash
-docker pull ghcr.io/lpalbou/abstractgateway-server:0.2.8
+docker pull ghcr.io/lpalbou/abstractgateway-server:0.2.15
 ```
 
 NVIDIA hosts can try the experimental full GPU image when local
 vLLM/HuggingFace/Diffusers engines are wanted. This image is published
 best-effort until it has a real CUDA build and smoke gate:
 
 ```bash
-docker pull ghcr.io/lpalbou/abstractgateway-server-nvidia:0.2.8
+docker pull ghcr.io/lpalbou/abstractgateway-server-nvidia:0.2.15
 ```
 
 The image installs the base `abstractgateway` package: HTTP server,
-`AbstractRuntime[multimodal]`, AbstractCore remote/commercial provider support,
-OpenAI-compatible text providers, workflow-backed and direct image generation
-through Runtime/Core/AbstractVision, direct Gateway voice/audio endpoints
-through AbstractVoice, provider/session prompt-cache helpers, media/tool
-helpers, token counting, compression, AbstractMemory/LanceDB KG support,
-AbstractAgent, and AbstractFlow compatibility.
+`AbstractRuntime[multimodal,mcp-worker]`, Runtime-owned provider/media/tool
+facades, OpenAI-compatible text providers, workflow-backed and direct image,
+voice, and audio routes surfaced through Runtime, provider/session prompt-cache
+helpers, AbstractMemory/LanceDB KG support, AbstractAgent, and AbstractFlow
+compatibility.
 
 ```bash
 export ABSTRACTGATEWAY_AUTH_TOKEN="$(python -c 'import secrets; print(secrets.token_urlsafe(32))')"
@@ -86,7 +85,7 @@ docker run --rm --name abstractgateway-server \
   -e OPENAI_COMPATIBLE_BASE_URL="http://host.docker.internal:1234/v1" \
   -v "$PWD/runtime/gateway:/data/gateway" \
   -v "$PWD/flows/bundles:/data/flows:ro" \
-  ghcr.io/lpalbou/abstractgateway-server:0.2.8
+  ghcr.io/lpalbou/abstractgateway-server:0.2.15
 ```
 
 On Apple Silicon, keep Metal/MLX inference native on macOS and run the
@@ -103,7 +102,7 @@ For a minimal Apple-local Gateway + Flow setup, see
 
 Compose and deployment details: [docs/deployment.md](docs/deployment.md).
 
-## 0.2.8 capability scope
+## 0.2.15 capability scope
 
 Direct Gateway APIs in this release:
 - `GET /api/gateway/runs/{run_id}/input_data`
@@ -122,9 +121,9 @@ Direct Gateway APIs in this release:
 - `/api/gateway/discovery/capabilities` package, plugin, and thin-client contract discovery
 
 Workflow/Core-backed capabilities:
-- Generated images are available to Runtime/Core workflows through
-  AbstractVision when installed and configured, and the direct Gateway image
-  route uses the same Runtime/Core image-generation contract.
+- Generated images are available to Runtime workflows through Runtime's image
+  backend integrations, and the direct Gateway image route uses the same
+  Runtime/Core image-generation contract.
 - Prompt-cache support depends on the active provider/model. Session lifecycle
   routes provide Gateway-owned naming and orchestration, not a provider-
   independent local KV cache.
@@ -148,8 +147,8 @@ See [docs/api.md](docs/api.md) for curl examples and the live OpenAPI spec (`/op
 Requires Python `>=3.10` (see `pyproject.toml`).
 
 The base install is the remote-light HTTP/SSE server: Gateway, Runtime,
-Agent, Core remote provider/tool/media support, Flow compatibility, Vision,
-Voice, and LanceDB-backed Memory.
+Agent, Flow compatibility, Runtime-owned provider/tool/media facades, and
+LanceDB-backed Memory.
 
 ```bash
 pip install abstractgateway

docker/abstractgateway-server/README.md

@@ -4,15 +4,15 @@ This image packages the AbstractGateway HTTP/SSE server for durable
 AbstractRuntime runs:
 
 ```bash
-ghcr.io/lpalbou/abstractgateway-server:0.2.8
+ghcr.io/lpalbou/abstractgateway-server:0.2.15
 ```
 
 Release images are published for `linux/amd64` and `linux/arm64`.
 A separate experimental full NVIDIA image is published best-effort for
 `linux/amd64`:
 
 ```bash
-ghcr.io/lpalbou/abstractgateway-server-nvidia:0.2.8
+ghcr.io/lpalbou/abstractgateway-server-nvidia:0.2.15
 ```
 
 The image installs:
@@ -22,13 +22,12 @@ abstractgateway==<version>
 ```
 
 The default image uses the base remote-light server install. It includes
-`AbstractRuntime[multimodal]`, AbstractCore
-remote/commercial provider support, OpenAI-compatible text provider routing,
-workflow-backed and direct image generation through Runtime/Core/AbstractVision,
-direct Gateway voice/audio endpoints through AbstractVoice, provider/session
-prompt-cache helpers, KG memory via AbstractMemory/LanceDB, media parsing,
-tool helpers, token counting, compression helpers, FastAPI/Uvicorn,
-AbstractAgent, and AbstractFlow compatibility. It
+`AbstractRuntime[multimodal,mcp-worker]`, Runtime-owned provider/media/tool
+facades, OpenAI-compatible text provider routing, workflow-backed and direct
+image generation through Runtime, direct Gateway voice/audio endpoints,
+provider/session prompt-cache helpers, KG memory via AbstractMemory/LanceDB,
+media parsing, tool helpers, token counting, compression helpers,
+FastAPI/Uvicorn, AbstractAgent, and AbstractFlow compatibility. It
 intentionally does not bundle local model runtimes such as MLX, vLLM,
 HuggingFace Transformers, local Diffusers/sdcpp vision backends, or local
 voice/music generation engines.
@@ -78,7 +77,7 @@ docker run --rm --name abstractgateway-server \
   -v "$PWD/runtime/gateway:/data/gateway" \
   -v "$PWD/flows/bundles:/data/flows:ro" \
   -v "$PWD/workspace:/workspace" \
-  ghcr.io/lpalbou/abstractgateway-server:0.2.8
+  ghcr.io/lpalbou/abstractgateway-server:0.2.15
 ```
 
 `ABSTRACTGATEWAY_AUTH_TOKEN` is the gateway bearer token. Clients send it as
@@ -97,7 +96,7 @@ docker run --rm --name abstractgateway-server \
   -e ABSTRACTGATEWAY_MODEL="your-model" \
   -e OPENAI_COMPATIBLE_BASE_URL="http://host.docker.internal:1234/v1" \
   -e OPENAI_COMPATIBLE_API_KEY="$OPENAI_COMPATIBLE_API_KEY" \
-  ghcr.io/lpalbou/abstractgateway-server:0.2.8
+  ghcr.io/lpalbou/abstractgateway-server:0.2.15
 ```
 
 Use `http://model-runner.docker.internal/engines/v1` for Docker Model Runner,
@@ -134,17 +133,17 @@ Useful compose variables:
 - `ABSTRACTVISION_*`: AbstractVision image backend or OpenAI-compatible image endpoint
 - `ABSTRACTVOICE_*`: AbstractVoice TTS/STT backend, local/remote engine, and model controls
 - `ABSTRACTGATEWAY_EXTRAS`: build-time install extra for local image builds (empty for the default image, `gpu` for NVIDIA)
-- `ABSTRACTGATEWAY_ABSTRACTCORE_SERVER_BASE_URL`: optional standalone Core server for dynamic voice/vision catalogs
+- `ABSTRACTCORE_SERVER_BASE_URL`: optional standalone Core server for dynamic voice/vision catalogs
 
 Release scope: TTS, STT, and generated images are direct Gateway endpoints.
-Generated images are also available through Runtime/Core workflows with
-AbstractVision installed and configured.
+Generated images are also available through Runtime workflows with a configured
+image backend.
 
 For unreleased local checkouts, build the image from this repository:
 
 ```bash
 ABSTRACTGATEWAY_INSTALL_MODE=local \
-ABSTRACTGATEWAY_IMAGE_TAG=0.2.8-local \
+ABSTRACTGATEWAY_IMAGE_TAG=0.2.15-local \
 docker compose -f docker/abstractgateway-server/compose.yml up -d --build
 ```
 

docker/abstractgateway-server/compose.nvidia.yml

@@ -2,13 +2,13 @@ services:
   abstractgateway:
     # Experimental CUDA overlay. Keep release-grade deployments on compose.yml
     # until this image has a real CUDA build/smoke gate.
-    image: ghcr.io/lpalbou/abstractgateway-server-nvidia:${ABSTRACTGATEWAY_IMAGE_TAG:-0.2.8}
+    image: ghcr.io/lpalbou/abstractgateway-server-nvidia:${ABSTRACTGATEWAY_IMAGE_TAG:-0.2.15}
     build:
       context: ../..
       dockerfile: docker/abstractgateway-server/Dockerfile.nvidia
       args:
         ABSTRACTGATEWAY_INSTALL_MODE: ${ABSTRACTGATEWAY_INSTALL_MODE:-pypi}
-        ABSTRACTGATEWAY_VERSION: ${ABSTRACTGATEWAY_VERSION:-0.2.8}
+        ABSTRACTGATEWAY_VERSION: ${ABSTRACTGATEWAY_VERSION:-0.2.15}
         ABSTRACTGATEWAY_EXTRAS: ${ABSTRACTGATEWAY_EXTRAS:-gpu}
     gpus: all
     environment:

docker/abstractgateway-server/compose.yml

@@ -2,13 +2,13 @@ name: abstractgateway-server
 
 services:
   abstractgateway:
-    image: ghcr.io/lpalbou/abstractgateway-server:${ABSTRACTGATEWAY_IMAGE_TAG:-0.2.8}
+    image: ghcr.io/lpalbou/abstractgateway-server:${ABSTRACTGATEWAY_IMAGE_TAG:-0.2.15}
     build:
       context: ../..
       dockerfile: docker/abstractgateway-server/Dockerfile
       args:
         ABSTRACTGATEWAY_INSTALL_MODE: ${ABSTRACTGATEWAY_INSTALL_MODE:-pypi}
-        ABSTRACTGATEWAY_VERSION: ${ABSTRACTGATEWAY_VERSION:-0.2.8}
+        ABSTRACTGATEWAY_VERSION: ${ABSTRACTGATEWAY_VERSION:-0.2.15}
         ABSTRACTGATEWAY_EXTRAS: ${ABSTRACTGATEWAY_EXTRAS:-}
     ports:
       - "${ABSTRACTGATEWAY_BIND_HOST:-127.0.0.1}:${ABSTRACTGATEWAY_PORT:-8080}:8080"
@@ -31,7 +31,7 @@ services:
       ABSTRACTGATEWAY_EMBEDDING_PROVIDER: ${ABSTRACTGATEWAY_EMBEDDING_PROVIDER:-disabled}
       ABSTRACTGATEWAY_EMBEDDING_MODEL: ${ABSTRACTGATEWAY_EMBEDDING_MODEL:-}
       ABSTRACTGATEWAY_WORKSPACE_DIR: /workspace
-      ABSTRACTGATEWAY_ABSTRACTCORE_SERVER_BASE_URL: ${ABSTRACTGATEWAY_ABSTRACTCORE_SERVER_BASE_URL:-}
+      ABSTRACTCORE_SERVER_BASE_URL: ${ABSTRACTCORE_SERVER_BASE_URL:-}
       ABSTRACTGATEWAY_ABSTRACTCORE_SERVER_AUTH_TOKEN: ${ABSTRACTGATEWAY_ABSTRACTCORE_SERVER_AUTH_TOKEN:-}
       ABSTRACTGATEWAY_CORE_CATALOG_TIMEOUT_S: ${ABSTRACTGATEWAY_CORE_CATALOG_TIMEOUT_S:-}
       OPENAI_API_KEY: ${OPENAI_API_KEY:-}

docs/README.md

@@ -7,7 +7,7 @@ Start here if you’re new to the project.
 AbstractGateway is one component in the larger AbstractFramework ecosystem:
 
 - **AbstractRuntime** (required): durable runs + stores
-- **AbstractCore / AbstractVoice / AbstractVision / AbstractMemory** (required by the default server install): LLM/tool execution, provider-level prompt-cache controls, workflow-backed/direct generated image/voice/audio capabilities, and KG memory used by many bundles
+- **AbstractRuntime + transitive capability packages** (required by the default server install): Runtime owns the LLM/tool/media integration boundary; Gateway uses its discovery/run facades for prompt-cache controls, generated image/voice/audio capabilities, and KG-backed bundle execution
 - Higher-level UIs (optional): AbstractFlow / AbstractObserver / AbstractCode / thin clients
 
 Related repos:

docs/api.md

@@ -260,14 +260,14 @@ Direct Gateway endpoints in this release:
 - `GET /api/gateway/vision/provider_models`
 
 The catalog endpoints proxy AbstractCore Server routes when
-`ABSTRACTGATEWAY_ABSTRACTCORE_SERVER_BASE_URL` or `ABSTRACTCORE_SERVER_BASE_URL`
+`ABSTRACTCORE_SERVER_BASE_URL`
 is configured. Gateway uses explicit Core auth settings for that hop and never
 reuses the Gateway bearer token as a Core/provider secret. Without a configured
 Core server, the voice/model routes return bounded static descriptors from
 Gateway and capability-package environment variables.
 
-Generated images are available through Runtime/Core workflows when
-AbstractVision is installed and configured. Gateway also exposes a direct image
+Generated images are available through Runtime workflows when a compatible
+image backend is installed and configured. Gateway also exposes a direct image
 generation endpoint that uses the Runtime/Core output-selector contract rather
 than a provider-specific image client. The route stores the generated image as a
 run artifact and emits `abstract.media.image.generated` with:
@@ -339,6 +339,31 @@ These routes derive a deterministic bounded namespace/key from `session_id`,
 provider operation results or a key hint. `rebuild` is clear-plus-prepare for
 providers that expose clear controls.
 
+Durable bloc exact-reuse endpoints:
+
+- `POST /api/gateway/blocs/upsert_text`
+- `GET /api/gateway/blocs/record`
+- `GET /api/gateway/blocs`
+- `POST /api/gateway/blocs/delete`
+- `GET /api/gateway/blocs/kv/manifest`
+- `GET /api/gateway/blocs/kv/list`
+- `POST /api/gateway/blocs/kv/ensure`
+- `POST /api/gateway/blocs/kv/load`
+- `POST /api/gateway/blocs/kv/delete`
+- `POST /api/gateway/blocs/kv/prune`
+
+These routes are the primary app-facing durable prompt-cache path:
+
+- create or identify a durable text bloc;
+- ensure or load a KV artifact for a target local provider/model;
+- use the returned `prompt_cache_binding` in later Runtime-backed generation;
+- list/delete/prune artifacts without reaching into provider-private cache state.
+
+They delegate through Runtime's public AbstractCore host facade rather than
+proxying Core directly. They are operator-style host controls, so the routes
+themselves are not ledgered run execution; the ledgered exact-reuse path is the
+later `LLM_CALL.params.prompt_cache_binding` used inside real Runtime runs.
+
 Provider-specific persistence endpoints:
 
 - `GET /api/gateway/prompt_cache/saved`
@@ -386,6 +411,6 @@ Configuration notes (gateway host):
 - Single-account env fallback: set `ABSTRACT_EMAIL_IMAP_*` and/or `ABSTRACT_EMAIL_SMTP_*`.
 - The secret itself must be present in the env var referenced by `*_PASSWORD_ENV_VAR` (e.g. `EMAIL_PASSWORD=...`).
 
-Evidence: `src/abstractgateway/routes/gateway.py` (`/email/accounts|messages|send`) which proxies to `abstractcore.tools.comms_tools`.
+Evidence: `src/abstractgateway/routes/gateway.py` (`/email/accounts|messages|send`) which proxies to the Runtime AbstractCore comms facade.
 
 Troubleshooting and common questions: [faq.md](./faq.md).

docs/architecture.md

@@ -1,7 +1,7 @@
 # AbstractGateway — Architecture
 
-> Status: implemented (v0.2.8)
-> Last reviewed: 2026-05-10
+> Status: implemented (v0.2.15)
+> Last reviewed: 2026-05-21
 
 AbstractGateway is a **durable run gateway** for AbstractRuntime:
 - **Start runs** (and optionally schedule them)
@@ -15,7 +15,7 @@ This document describes the code in this repository (see **Evidence** links).
 AbstractGateway is designed to sit between **thin clients / UIs** and **AbstractRuntime**:
 - AbstractGateway: HTTP/SSE API + durability glue + baseline security (`src/abstractgateway/app.py`, `src/abstractgateway/routes/gateway.py`)
 - AbstractRuntime (required): run model + tick loop + stores (`pyproject.toml`, `src/abstractgateway/runner.py`)
-- AbstractCore / AbstractVoice / AbstractVision / AbstractMemory (required by the default server install): LLM/tool execution, provider-level prompt-cache controls, workflow-backed/direct generated image/voice/audio capabilities, and KG memory used by many bundles (`src/abstractgateway/hosts/bundle_host.py`)
+- AbstractRuntime + transitive capability packages (required by the default server install): Runtime owns the LLM/tool/media integration boundary; Gateway uses its discovery/run facades for prompt-cache controls, generated image/voice/audio capabilities, and KG-backed bundle execution (`src/abstractgateway/hosts/bundle_host.py`)
 
 ## High-level shape