diff --git a/CHANGELOG.md b/CHANGELOG.md
index 8f5fcbe..2ce7cad 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -1,5 +1,51 @@
 # Changelog
 
+## 0.5.0 — 2026-07-06
+
+### Added
+- **Media endpoints on the FastAPI sidecar** (#14). The proxy only exposed
+  `/v1/images/generations`, so OpenAI-compatible clients (LiteLLM) could not
+  reach the gateway's video/audio models. Adds:
+  - `POST /v1/videos/generations`
+  - `POST /v1/audio/speech`
+  - `POST /v1/audio/generations` (music)
+  - `POST /v1/audio/sound-effects`
+
+  The adapter dispatches Base (dedicated `VideoClient`/`MusicClient`/
+  `SpeechClient`) vs Solana (unified `SolanaLLMClient`) per request; sync SDK
+  clients run in a shared thread pool. Validated live end-to-end on Solana
+  mainnet.
+
+### Changed
+- **Require `blockrun-llm[solana]>=1.5.0`** for the `solana` extra — Solana
+  media (`SolanaLLMClient.video/music/speech/sound_effect`) landed in the SDK's
+  1.5.0 release (blockrun-llm #16; also carries #17 rpc_batch cache fix, #18
+  `solana<0.40` pin, #19 media hardening). Older SDKs degrade Solana media to a
+  clear 501, not a 500. The core `blockrun-llm>=1.4.7` floor is unchanged — Base
+  media clients predate it.
+- Synced `__version__` (was stale at 0.4.2).
+
+### Hardened (#15, media-endpoints review follow-up)
+- `ValueError → 400` on **all** media routes via a shared `_media_endpoint`
+  helper (was video-only; music-with-lyrics 500'd instead of returning the SDK's
+  clear message).
+- Solana media on a pre-1.5.0 SDK degrades to a clear **501** upgrade hint
+  instead of an `AttributeError` 500.
+- Long media (video 60–900s, music 60–210s) moved to a dedicated 8-thread pool
+  (`BLOCKRUN_LONG_MEDIA_THREADS`); all media routes gated behind their own
+  semaphore (`BLOCKRUN_MEDIA_MAX_CONCURRENT`, default 20) so a video burst can no
+  longer starve images or brick chat/messages.
+- Client-supplied `budget_seconds`/`timeout` clamped to the 900s server cap
+  (was forwarded verbatim — one request body could pin a worker for a day);
+  every media call wrapped in `asyncio.wait_for` so the coroutine + semaphore
+  permit always release even if the SDK thread wedges (504).
+- Media calls now audit-log via `log_proxy_call` and surface the in-body
+  settlement txHash as `x-blockrun-settlement` (paid media traffic was invisible
+  to spend reconciliation). NB: media SDK responses don't expose `cost_usd` yet,
+  so media audit rows log `cost_usd=None` with only the settlement txHash.
+- Full negative-path pytest suites for all 5 media routes + adapter
+  dispatch/clamp/guard/ceiling tests.
+
 ## 0.4.7 — 2026-06-26
 
 ### Added
diff --git a/blockrun_litellm/__init__.py b/blockrun_litellm/__init__.py
index 6e363a9..1325193 100644
--- a/blockrun_litellm/__init__.py
+++ b/blockrun_litellm/__init__.py
@@ -26,4 +26,4 @@
 from blockrun_litellm.provider import BlockRunLLM, register
 
 __all__ = ["BlockRunLLM", "register", "enable_local_logging"]
-__version__ = "0.4.2"
+__version__ = "0.5.0"
diff --git a/pyproject.toml b/pyproject.toml
index 6ed6376..d310eb8 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "blockrun-litellm"
-version = "0.4.7"
+version = "0.5.0"
 description = "LiteLLM adapter for BlockRun — call x402-paid AI models via LiteLLM (custom provider or local OpenAI-compatible proxy)"
 readme = "README.md"
 license = "MIT"
@@ -40,9 +40,10 @@ dependencies = [
 # Re-exported here so `pip install 'blockrun-litellm[solana]'` pulls the
 # x402 SVM toolchain that SolanaLLMClient needs (x402 SDK + solders).
 # NB: Solana *media* (SolanaLLMClient.video/music/speech/sound_effect) needs the
-# blockrun-llm release carrying BlockRunAI/blockrun-llm#16 — bump this floor when
-# it ships. Until then the adapter degrades those calls to a clear 501, not a 500.
-solana = ["blockrun-llm[solana]>=1.4.7"]
+# blockrun-llm release carrying BlockRunAI/blockrun-llm#16, shipped in 1.5.0
+# (also carries #17 rpc_batch cache fix, #18 solana<0.40 pin, #19 media
+# hardening). Older SDKs degrade Solana media to a clear 501, not a 500.
+solana = ["blockrun-llm[solana]>=1.5.0"]
 proxy = [
     "fastapi>=0.110.0",
     "uvicorn[standard]>=0.27.0",