diff --git a/services/planner-agent/requirements.txt b/services/planner-agent/requirements.txt
new file mode 100644
index 0000000..3d88ca9
--- /dev/null
+++ b/services/planner-agent/requirements.txt
@@ -0,0 +1,3 @@
+litellm>=1.40.0
+redis[asyncio]>=5.0.0
+jsonschema>=4.21.0
diff --git a/services/planner-agent/src/llm_router.py b/services/planner-agent/src/llm_router.py
new file mode 100644
index 0000000..505a89b
--- /dev/null
+++ b/services/planner-agent/src/llm_router.py
@@ -0,0 +1,447 @@
+"""
+llm_router.py — LLM routing with local-first fallback chain.
+
+Routing strategy:
+  1. Local Qwen via Ollama   (piha:11434, timeout 8 s)
+  2. claude-haiku-4-5        (Anthropic cloud, timeout 30 s)
+  3. claude-sonnet-4-6       (Anthropic cloud, timeout 30 s)
+
+A model is rejected when it:
+  - times out
+  - raises any API / network exception
+  - returns text matching a refusal pattern
+  - returns JSON that fails the caller-supplied JSON Schema
+
+After every call (success or full-chain failure) a metrics event is
+published to the Redis channel "llm_router_metrics".
+
+Usage
+-----
+    router = LLMRouter()
+    result = await router.route(
+        messages=[{"role": "user", "content": "What should I do?"}],
+        schema={"type": "object", "required": ["action"], "properties": {...}},
+    )
+    print(result.model_used, result.content)
+    await router.close()
+"""
+
+import asyncio
+import json
+import logging
+import time
+from dataclasses import dataclass, field
+from typing import Any, Optional
+
+import litellm
+import redis.asyncio as aioredis
+from jsonschema import validate, ValidationError
+
+litellm.suppress_debug_info = True
+
+logger = logging.getLogger("llm_router")
+
+# ---------------------------------------------------------------------------
+# Refusal patterns — any substring match (case-insensitive) triggers fallback
+# ---------------------------------------------------------------------------
+REFUSAL_PATTERNS: list[str] = [
+    "nie wiem",
+    "I cannot",
+    "I can't",
+    "as an AI",
+    "I don't know",
+    "I'm not able",
+    "I am not able",
+    "I'm unable",
+    "I am unable",
+    "beyond my capabilities",
+]
+
+
+# ---------------------------------------------------------------------------
+# Data structures
+# ---------------------------------------------------------------------------
+
+@dataclass
+class ModelConfig:
+    """Configuration for one model in the fallback chain."""
+    name: str                          # litellm model string, e.g. "ollama/qwen2.5:7b"
+    timeout: float                     # hard wall-clock timeout in seconds
+    api_base: Optional[str] = None     # override API base URL (Ollama needs this)
+    extra_kwargs: dict = field(default_factory=dict)
+
+    def __str__(self) -> str:
+        base = f" @ {self.api_base}" if self.api_base else ""
+        return f"{self.name}{base} (timeout={self.timeout}s)"
+
+
+@dataclass
+class AttemptRecord:
+    model: str
+    outcome: str          # "success" | "rejected" | "invalid"
+    reason: Optional[str] # None on success
+    latency_ms: int
+
+
+@dataclass
+class RouteResult:
+    """Return value of LLMRouter.route()."""
+    content: Any                        # parsed JSON (if schema given) or raw str
+    raw_text: str
+    model_used: str                     # "none" if every model failed
+    attempts: list[AttemptRecord]
+    latency_ms: int                     # wall-clock from first attempt to return
+
+    @property
+    def succeeded(self) -> bool:
+        return self.model_used != "none"
+
+    def to_dict(self) -> dict:
+        return {
+            "model_used": self.model_used,
+            "latency_ms": self.latency_ms,
+            "attempts": [
+                {
+                    "model":      a.model,
+                    "outcome":    a.outcome,
+                    "reason":     a.reason,
+                    "latency_ms": a.latency_ms,
+                }
+                for a in self.attempts
+            ],
+        }
+
+
+# ---------------------------------------------------------------------------
+# Metrics
+# ---------------------------------------------------------------------------
+
+class ModelMetrics:
+    """Thread-safe-ish counter per model × outcome.
+
+    Outcomes: "success", "fallback", "error"
+    ("fallback" = rejected but another model succeeded after it;
+     "error"    = rejected and it was the last in chain or chain exhausted)
+    """
+
+    def __init__(self) -> None:
+        self._counts: dict[str, dict[str, int]] = {}
+
+    def record(self, model: str, outcome: str) -> None:
+        if model not in self._counts:
+            self._counts[model] = {"success": 0, "fallback": 0, "error": 0}
+        self._counts[model][outcome] = self._counts[model].get(outcome, 0) + 1
+
+    def snapshot(self) -> dict[str, dict[str, int]]:
+        return {m: dict(c) for m, c in self._counts.items()}
+
+    def total_calls(self, model: str) -> int:
+        return sum(self._counts.get(model, {}).values())
+
+    def success_rate(self, model: str) -> Optional[float]:
+        counts = self._counts.get(model, {})
+        total = sum(counts.values())
+        if total == 0:
+            return None
+        return counts.get("success", 0) / total
+
+
+# ---------------------------------------------------------------------------
+# Router
+# ---------------------------------------------------------------------------
+
+class LLMRouter:
+    """Route LLM calls through a local-first fallback chain.
+
+    Parameters
+    ----------
+    redis_url:
+        Redis connection URL for metrics publishing.
+        Set to None to disable Redis (useful in tests / local dev).
+    ollama_host:
+        Base URL of the Ollama API.  Defaults to piha's Tailscale address.
+    ollama_model:
+        Model tag as known to Ollama (e.g. "qwen2.5:7b").
+    chain:
+        Override the entire fallback chain.  When None the default
+        Qwen → haiku → sonnet chain is used.
+    """
+
+    DEFAULT_OLLAMA_HOST  = "http://100.108.208.3:11434"
+    DEFAULT_OLLAMA_MODEL = "qwen2.5:7b"
+    DEFAULT_REDIS_URL    = "redis://100.108.208.3:6379"
+
+    def __init__(
+        self,
+        redis_url: Optional[str] = DEFAULT_REDIS_URL,
+        ollama_host: str = DEFAULT_OLLAMA_HOST,
+        ollama_model: str = DEFAULT_OLLAMA_MODEL,
+        chain: Optional[list[ModelConfig]] = None,
+    ) -> None:
+        if chain is not None:
+            self.chain = chain
+        else:
+            self.chain = [
+                ModelConfig(
+                    name=f"ollama/{ollama_model}",
+                    timeout=8.0,
+                    api_base=ollama_host,
+                ),
+                ModelConfig(
+                    name="claude-haiku-4-5-20251001",
+                    timeout=30.0,
+                ),
+                ModelConfig(
+                    name="claude-sonnet-4-6",
+                    timeout=30.0,
+                ),
+            ]
+
+        self.metrics = ModelMetrics()
+        self._redis_url = redis_url
+        self._redis: Optional[aioredis.Redis] = None
+
+    # ------------------------------------------------------------------
+    # Public API
+    # ------------------------------------------------------------------
+
+    async def route(
+        self,
+        messages: list[dict],
+        schema: Optional[dict] = None,
+        context: Optional[str] = None,
+    ) -> RouteResult:
+        """Try each model in order; return the first valid response.
+
+        Parameters
+        ----------
+        messages:
+            OpenAI-style message list, e.g.
+            [{"role": "user", "content": "..."}]
+        schema:
+            Optional JSON Schema dict.  When provided the model's response
+            must be valid JSON that conforms to the schema.
+        context:
+            Optional free-text caller label included in log lines (e.g.
+            "supervisor.reconcile") for easier tracing.
+
+        Raises
+        ------
+        RuntimeError
+            When every model in the chain fails.  The exception message
+            contains a JSON-formatted attempt log.
+        """
+        tag = f"[{context}] " if context else ""
+        start = time.monotonic()
+        attempts: list[AttemptRecord] = []
+
+        for i, cfg in enumerate(self.chain):
+            is_last = i == len(self.chain) - 1
+            attempt_start = time.monotonic()
+
+            logger.info(
+                f"{tag}[llm_router] attempt {i+1}/{len(self.chain)}: {cfg}"
+            )
+
+            raw_text, call_error = await self._call_model(cfg, messages)
+            attempt_ms = round((time.monotonic() - attempt_start) * 1000)
+
+            if call_error:
+                self.metrics.record(cfg.name, "error" if is_last else "fallback")
+                logger.warning(
+                    f"{tag}[llm_router] {cfg.name} → rejected "
+                    f"({call_error}) [{attempt_ms}ms]"
+                )
+                attempts.append(AttemptRecord(
+                    model=cfg.name, outcome="rejected",
+                    reason=call_error, latency_ms=attempt_ms,
+                ))
+                continue
+
+            parsed, schema_error = self._validate(raw_text, schema)
+            if schema_error:
+                self.metrics.record(cfg.name, "error" if is_last else "fallback")
+                logger.warning(
+                    f"{tag}[llm_router] {cfg.name} → invalid "
+                    f"({schema_error}) [{attempt_ms}ms]"
+                )
+                attempts.append(AttemptRecord(
+                    model=cfg.name, outcome="invalid",
+                    reason=schema_error, latency_ms=attempt_ms,
+                ))
+                continue
+
+            # ── success ───────────────────────────────────────────────
+            self.metrics.record(cfg.name, "success")
+            total_ms = round((time.monotonic() - start) * 1000)
+            logger.info(
+                f"{tag}[llm_router] {cfg.name} → success "
+                f"[attempt {attempt_ms}ms, total {total_ms}ms]"
+            )
+            attempts.append(AttemptRecord(
+                model=cfg.name, outcome="success",
+                reason=None, latency_ms=attempt_ms,
+            ))
+            result = RouteResult(
+                content=parsed,
+                raw_text=raw_text,
+                model_used=cfg.name,
+                attempts=attempts,
+                latency_ms=total_ms,
+            )
+            await self._publish_metrics(result)
+            return result
+
+        # ── all models exhausted ──────────────────────────────────────
+        total_ms = round((time.monotonic() - start) * 1000)
+        result = RouteResult(
+            content=None,
+            raw_text="",
+            model_used="none",
+            attempts=attempts,
+            latency_ms=total_ms,
+        )
+        await self._publish_metrics(result)
+
+        attempt_log = json.dumps(
+            [{"model": a.model, "reason": a.reason} for a in attempts],
+            indent=2,
+        )
+        raise RuntimeError(
+            f"{tag}[llm_router] All {len(self.chain)} models in chain failed.\n"
+            f"Attempts:\n{attempt_log}"
+        )
+
+    async def close(self) -> None:
+        """Release the Redis connection."""
+        if self._redis is not None:
+            await self._redis.aclose()
+            self._redis = None
+
+    # ------------------------------------------------------------------
+    # Internal helpers
+    # ------------------------------------------------------------------
+
+    async def _call_model(
+        self,
+        cfg: ModelConfig,
+        messages: list[dict],
+    ) -> tuple[str, Optional[str]]:
+        """Invoke one model.  Returns (raw_text, error_reason|None)."""
+        kwargs: dict[str, Any] = {
+            "model":    cfg.name,
+            "messages": messages,
+            "timeout":  cfg.timeout,
+            **cfg.extra_kwargs,
+        }
+        if cfg.api_base:
+            kwargs["api_base"] = cfg.api_base
+
+        try:
+            # asyncio.wait_for as belt-and-suspenders — litellm timeout
+            # is passed to the underlying HTTP client, but asyncio task
+            # cancellation ensures we never block the event loop.
+            resp = await asyncio.wait_for(
+                litellm.acompletion(**kwargs),
+                timeout=cfg.timeout + 2,   # +2 s grace for HTTP overhead
+            )
+            text = (resp.choices[0].message.content or "").strip()
+        except asyncio.TimeoutError:
+            return "", f"Timeout after {cfg.timeout}s"
+        except litellm.exceptions.Timeout:
+            return "", f"Timeout after {cfg.timeout}s"
+        except litellm.exceptions.APIConnectionError as e:
+            return "", f"APIConnectionError: {e}"
+        except litellm.exceptions.AuthenticationError as e:
+            return "", f"AuthenticationError: {e}"
+        except Exception as e:
+            return "", f"{type(e).__name__}: {e}"
+
+        # Check for refusals in the model's own text
+        refusal = self._detect_refusal(text)
+        if refusal:
+            return text, f"RefusalPattern matched: '{refusal}'"
+        return text, None
+
+    @staticmethod
+    def _detect_refusal(text: str) -> Optional[str]:
+        """Return the first matching refusal pattern, or None."""
+        lower = text.lower()
+        for pattern in REFUSAL_PATTERNS:
+            if pattern.lower() in lower:
+                return pattern
+        return None
+
+    @staticmethod
+    def _validate(
+        text: str,
+        schema: Optional[dict],
+    ) -> tuple[Any, Optional[str]]:
+        """Parse and validate the model response.
+
+        Returns (parsed_content, error_reason|None).
+        When schema is None, returns (raw_text, None) — only refusal
+        detection (already done in _call_model) applies.
+        """
+        if schema is None:
+            return text, None
+
+        try:
+            parsed = json.loads(text)
+        except json.JSONDecodeError as exc:
+            # Try to extract JSON from a markdown code fence
+            extracted = _extract_json_from_fence(text)
+            if extracted is not None:
+                parsed = extracted
+            else:
+                return None, f"JSONDecodeError: {exc}"
+
+        try:
+            validate(instance=parsed, schema=schema)
+        except ValidationError as exc:
+            return None, f"SchemaValidationError: {exc.message}"
+
+        return parsed, None
+
+    async def _get_redis(self) -> Optional[aioredis.Redis]:
+        if self._redis_url is None:
+            return None
+        if self._redis is None:
+            self._redis = aioredis.from_url(
+                self._redis_url,
+                decode_responses=True,
+                socket_connect_timeout=2,
+                socket_timeout=2,
+            )
+        return self._redis
+
+    async def _publish_metrics(self, result: RouteResult) -> None:
+        """Non-blocking publish to Redis channel 'llm_router_metrics'."""
+        payload = {
+            **result.to_dict(),
+            "metrics_snapshot": self.metrics.snapshot(),
+            "timestamp": time.time(),
+        }
+        try:
+            r = await self._get_redis()
+            if r is not None:
+                await r.publish("llm_router_metrics", json.dumps(payload))
+        except Exception as exc:
+            # Never let a metrics failure break the caller
+            logger.warning(f"[llm_router] metrics publish failed: {exc}")
+
+
+# ---------------------------------------------------------------------------
+# Utility
+# ---------------------------------------------------------------------------
+
+def _extract_json_from_fence(text: str) -> Optional[Any]:
+    """Extract JSON from a ```json ... ``` markdown code fence, if present."""
+    import re
+    match = re.search(r"```(?:json)?\s*(\{.*?\}|\[.*?\])\s*```", text, re.DOTALL)
+    if match:
+        try:
+            return json.loads(match.group(1))
+        except json.JSONDecodeError:
+            pass
+    return None
diff --git a/services/planner-agent/tests/test_llm_router.py b/services/planner-agent/tests/test_llm_router.py
new file mode 100644
index 0000000..9c025de
--- /dev/null
+++ b/services/planner-agent/tests/test_llm_router.py
@@ -0,0 +1,449 @@
+"""
+Unit tests for llm_router.py.
+
+All LLM and Redis calls are mocked — no network required.
+
+Run:
+    pip install pytest pytest-asyncio litellm jsonschema redis
+    pytest services/planner-agent/tests/test_llm_router.py -v
+"""
+
+import asyncio
+import json
+import sys
+from pathlib import Path
+from typing import Any
+from unittest.mock import AsyncMock, MagicMock, patch
+
+import pytest
+
+# Allow importing from src/ without installation
+sys.path.insert(0, str(Path(__file__).parent.parent / "src"))
+
+from llm_router import (
+    AttemptRecord,
+    LLMRouter,
+    ModelConfig,
+    ModelMetrics,
+    RouteResult,
+    _extract_json_from_fence,
+)
+
+# ---------------------------------------------------------------------------
+# Helpers
+# ---------------------------------------------------------------------------
+
+def _fake_completion(content: str):
+    """Build a minimal litellm-style response object."""
+    msg = MagicMock()
+    msg.content = content
+    choice = MagicMock()
+    choice.message = msg
+    resp = MagicMock()
+    resp.choices = [choice]
+    return resp
+
+
+def _chain_of(*models: tuple[str, float]) -> list[ModelConfig]:
+    """Build a minimal test chain with no api_base."""
+    return [ModelConfig(name=name, timeout=timeout) for name, timeout in models]
+
+
+# ---------------------------------------------------------------------------
+# ModelMetrics
+# ---------------------------------------------------------------------------
+
+class TestModelMetrics:
+    def test_record_and_snapshot(self):
+        m = ModelMetrics()
+        m.record("qwen", "success")
+        m.record("qwen", "success")
+        m.record("qwen", "fallback")
+        m.record("haiku", "success")
+
+        snap = m.snapshot()
+        assert snap["qwen"]["success"] == 2
+        assert snap["qwen"]["fallback"] == 1
+        assert snap["haiku"]["success"] == 1
+
+    def test_success_rate(self):
+        m = ModelMetrics()
+        m.record("model-a", "success")
+        m.record("model-a", "fallback")
+        assert m.success_rate("model-a") == 0.5
+
+    def test_success_rate_unknown_model(self):
+        m = ModelMetrics()
+        assert m.success_rate("ghost") is None
+
+    def test_total_calls(self):
+        m = ModelMetrics()
+        m.record("x", "success")
+        m.record("x", "error")
+        assert m.total_calls("x") == 2
+
+    def test_snapshot_is_copy(self):
+        m = ModelMetrics()
+        m.record("x", "success")
+        snap = m.snapshot()
+        snap["x"]["success"] = 999  # mutate the copy
+        assert m.snapshot()["x"]["success"] == 1  # original unchanged
+
+
+# ---------------------------------------------------------------------------
+# RouteResult
+# ---------------------------------------------------------------------------
+
+class TestRouteResult:
+    def test_succeeded(self):
+        r = RouteResult("hello", "hello", "model-a", [], 100)
+        assert r.succeeded is True
+
+    def test_not_succeeded(self):
+        r = RouteResult(None, "", "none", [], 100)
+        assert r.succeeded is False
+
+    def test_to_dict_structure(self):
+        attempt = AttemptRecord("m", "success", None, 50)
+        r = RouteResult("x", "x", "m", [attempt], 50)
+        d = r.to_dict()
+        assert d["model_used"] == "m"
+        assert len(d["attempts"]) == 1
+        assert d["attempts"][0]["outcome"] == "success"
+
+
+# ---------------------------------------------------------------------------
+# _extract_json_from_fence
+# ---------------------------------------------------------------------------
+
+class TestExtractJsonFromFence:
+    def test_json_fence(self):
+        text = 'Sure!\n```json\n{"a": 1}\n```\nDone.'
+        assert _extract_json_from_fence(text) == {"a": 1}
+
+    def test_plain_fence(self):
+        text = "```\n[1, 2]\n```"
+        assert _extract_json_from_fence(text) == [1, 2]
+
+    def test_no_fence(self):
+        assert _extract_json_from_fence("no json here") is None
+
+    def test_broken_fence(self):
+        assert _extract_json_from_fence("```json\n{broken```") is None
+
+
+# ---------------------------------------------------------------------------
+# LLMRouter — validation & refusal detection
+# ---------------------------------------------------------------------------
+
+class TestLLMRouterValidation:
+    def setup_method(self):
+        self.router = LLMRouter(redis_url=None)
+
+    def test_validate_no_schema_returns_text(self):
+        parsed, err = self.router._validate("hello world", schema=None)
+        assert parsed == "hello world"
+        assert err is None
+
+    def test_validate_valid_json(self):
+        schema = {"type": "object", "required": ["action"]}
+        parsed, err = self.router._validate('{"action": "redeploy"}', schema)
+        assert err is None
+        assert parsed == {"action": "redeploy"}
+
+    def test_validate_invalid_json(self):
+        schema = {"type": "object"}
+        _, err = self.router._validate("not json {", schema)
+        assert err is not None
+        assert "JSONDecodeError" in err
+
+    def test_validate_schema_violation(self):
+        schema = {"type": "object", "required": ["action"]}
+        _, err = self.router._validate('{"other": 1}', schema)
+        assert err is not None
+        assert "SchemaValidationError" in err
+
+    def test_validate_extracts_fenced_json(self):
+        schema = {"type": "object", "required": ["action"]}
+        text = '```json\n{"action": "restart"}\n```'
+        parsed, err = self.router._validate(text, schema)
+        assert err is None
+        assert parsed == {"action": "restart"}
+
+    def test_detect_refusal_nie_wiem(self):
+        assert self.router._detect_refusal("Nie wiem co mam zrobić") == "nie wiem"
+
+    def test_detect_refusal_as_an_ai(self):
+        # Text contains both "as an AI" and "I cannot"; first match wins.
+        # We only assert that a refusal IS detected, not which pattern fires.
+        assert self.router._detect_refusal("As an AI I cannot help") is not None
+
+    def test_detect_refusal_none(self):
+        assert self.router._detect_refusal("Sure, here is the action.") is None
+
+    def test_detect_refusal_case_insensitive(self):
+        assert self.router._detect_refusal("I CANNOT do that") == "I cannot"
+
+
+# ---------------------------------------------------------------------------
+# LLMRouter — routing logic (mocked litellm + Redis)
+# ---------------------------------------------------------------------------
+
+@pytest.fixture
+def router_no_redis():
+    """Router with a 3-model chain and Redis disabled."""
+    chain = _chain_of(
+        ("local/qwen", 8.0),
+        ("claude-haiku-test", 30.0),
+        ("claude-sonnet-test", 30.0),
+    )
+    return LLMRouter(redis_url=None, chain=chain)
+
+
+@pytest.mark.asyncio
+class TestLLMRouterRouting:
+    async def test_primary_success(self, router_no_redis):
+        """Primary model succeeds — no fallback."""
+        with patch(
+            "litellm.acompletion",
+            AsyncMock(return_value=_fake_completion('{"action": "ok"}')),
+        ):
+            result = await router_no_redis.route(
+                messages=[{"role": "user", "content": "test"}],
+                schema={"type": "object", "required": ["action"]},
+            )
+
+        assert result.model_used == "local/qwen"
+        assert result.content == {"action": "ok"}
+        assert len(result.attempts) == 1
+        assert result.attempts[0].outcome == "success"
+
+    async def test_fallback_on_json_error(self, router_no_redis):
+        """Primary returns bad JSON → falls back to haiku which returns valid JSON."""
+        responses = [
+            _fake_completion("not json at all"),
+            _fake_completion('{"action": "restart"}'),
+        ]
+        call_count = 0
+
+        async def fake_acompletion(**kwargs):
+            nonlocal call_count
+            r = responses[call_count]
+            call_count += 1
+            return r
+
+        with patch("litellm.acompletion", fake_acompletion):
+            result = await router_no_redis.route(
+                messages=[{"role": "user", "content": "x"}],
+                schema={"type": "object", "required": ["action"]},
+            )
+
+        assert result.model_used == "claude-haiku-test"
+        assert result.content == {"action": "restart"}
+        assert result.attempts[0].outcome == "invalid"
+        assert result.attempts[1].outcome == "success"
+
+    async def test_fallback_on_refusal(self, router_no_redis):
+        """Primary returns refusal text → falls back."""
+        responses = [
+            _fake_completion("I cannot help with that request."),
+            _fake_completion("Sure! Here is the plan."),
+        ]
+        idx = 0
+
+        async def fake_acompletion(**kwargs):
+            nonlocal idx
+            r = responses[idx]
+            idx += 1
+            return r
+
+        with patch("litellm.acompletion", fake_acompletion):
+            result = await router_no_redis.route(
+                messages=[{"role": "user", "content": "x"}],
+            )
+
+        assert result.model_used == "claude-haiku-test"
+        assert result.attempts[0].outcome == "rejected"
+        assert "RefusalPattern" in result.attempts[0].reason
+
+    async def test_fallback_on_timeout(self, router_no_redis):
+        """Primary times out → falls back to haiku."""
+        call_count = 0
+
+        async def fake_acompletion(**kwargs):
+            nonlocal call_count
+            call_count += 1
+            if call_count == 1:
+                raise asyncio.TimeoutError()
+            return _fake_completion("fallback response")
+
+        with patch("litellm.acompletion", fake_acompletion):
+            result = await router_no_redis.route(
+                messages=[{"role": "user", "content": "x"}],
+            )
+
+        assert result.model_used == "claude-haiku-test"
+        assert "Timeout" in result.attempts[0].reason
+
+    async def test_fallback_on_api_exception(self, router_no_redis):
+        """Primary raises a connection error → falls back."""
+        import litellm.exceptions
+
+        call_count = 0
+
+        async def fake_acompletion(**kwargs):
+            nonlocal call_count
+            call_count += 1
+            if call_count == 1:
+                raise litellm.exceptions.APIConnectionError(
+                    message="Connection refused", llm_provider="ollama", model="qwen"
+                )
+            return _fake_completion("ok")
+
+        with patch("litellm.acompletion", fake_acompletion):
+            result = await router_no_redis.route(
+                messages=[{"role": "user", "content": "x"}],
+            )
+
+        assert result.model_used == "claude-haiku-test"
+        assert "APIConnectionError" in result.attempts[0].reason
+
+    async def test_all_models_fail_raises(self, router_no_redis):
+        """All models return bad JSON → RuntimeError with attempt log."""
+        with patch(
+            "litellm.acompletion",
+            AsyncMock(return_value=_fake_completion("not json")),
+        ):
+            with pytest.raises(RuntimeError) as exc_info:
+                await router_no_redis.route(
+                    messages=[{"role": "user", "content": "x"}],
+                    schema={"type": "object"},
+                )
+
+        assert "All 3 models in chain failed" in str(exc_info.value)
+
+    async def test_schema_none_no_json_required(self, router_no_redis):
+        """Without a schema, plain text responses are accepted."""
+        with patch(
+            "litellm.acompletion",
+            AsyncMock(return_value=_fake_completion("Here is your plan.")),
+        ):
+            result = await router_no_redis.route(
+                messages=[{"role": "user", "content": "x"}],
+            )
+
+        assert result.content == "Here is your plan."
+        assert result.succeeded
+
+    async def test_metrics_recorded_on_success(self, router_no_redis):
+        with patch(
+            "litellm.acompletion",
+            AsyncMock(return_value=_fake_completion("ok")),
+        ):
+            await router_no_redis.route([{"role": "user", "content": "x"}])
+
+        snap = router_no_redis.metrics.snapshot()
+        assert snap["local/qwen"]["success"] == 1
+
+    async def test_metrics_fallback_recorded(self, router_no_redis):
+        """Primary fails → fallback → metrics show primary=fallback, haiku=success."""
+        responses = [
+            _fake_completion("I cannot help"),  # refusal
+            _fake_completion("ok"),
+        ]
+        idx = 0
+
+        async def fake_acompletion(**kwargs):
+            nonlocal idx
+            r = responses[idx]; idx += 1
+            return r
+
+        with patch("litellm.acompletion", fake_acompletion):
+            await router_no_redis.route([{"role": "user", "content": "x"}])
+
+        snap = router_no_redis.metrics.snapshot()
+        assert snap["local/qwen"]["fallback"] == 1
+        assert snap["claude-haiku-test"]["success"] == 1
+
+    async def test_context_label_in_logs(self, router_no_redis, caplog):
+        """context= parameter appears in log output."""
+        import logging
+        with patch(
+            "litellm.acompletion",
+            AsyncMock(return_value=_fake_completion("ok")),
+        ):
+            with caplog.at_level(logging.INFO, logger="llm_router"):
+                await router_no_redis.route(
+                    messages=[{"role": "user", "content": "x"}],
+                    context="supervisor.reconcile",
+                )
+
+        assert any("supervisor.reconcile" in r.message for r in caplog.records)
+
+
+# ---------------------------------------------------------------------------
+# LLMRouter — Redis metrics publish
+# ---------------------------------------------------------------------------
+
+@pytest.mark.asyncio
+class TestLLMRouterRedis:
+    async def test_metrics_published_on_success(self):
+        chain = _chain_of(("m1", 8.0),)
+        router = LLMRouter(redis_url="redis://localhost:6379", chain=chain)
+
+        mock_redis = AsyncMock()
+        router._redis = mock_redis
+
+        with patch(
+            "litellm.acompletion",
+            AsyncMock(return_value=_fake_completion("ok")),
+        ):
+            await router.route([{"role": "user", "content": "x"}])
+
+        mock_redis.publish.assert_awaited_once()
+        channel, payload_str = mock_redis.publish.call_args[0]
+        assert channel == "llm_router_metrics"
+        payload = json.loads(payload_str)
+        assert payload["model_used"] == "m1"
+        assert "metrics_snapshot" in payload
+        assert "timestamp" in payload
+
+    async def test_redis_failure_does_not_raise(self):
+        """A broken Redis must never break the LLM call result."""
+        chain = _chain_of(("m1", 8.0),)
+        router = LLMRouter(redis_url="redis://localhost:6379", chain=chain)
+
+        mock_redis = AsyncMock()
+        mock_redis.publish.side_effect = ConnectionError("Redis down")
+        router._redis = mock_redis
+
+        with patch(
+            "litellm.acompletion",
+            AsyncMock(return_value=_fake_completion("ok")),
+        ):
+            result = await router.route([{"role": "user", "content": "x"}])
+
+        assert result.succeeded  # LLM call still returned
+
+    async def test_metrics_published_on_full_failure(self):
+        chain = _chain_of(("m1", 8.0),)
+        router = LLMRouter(redis_url="redis://localhost:6379", chain=chain)
+
+        mock_redis = AsyncMock()
+        router._redis = mock_redis
+
+        with patch(
+            "litellm.acompletion",
+            AsyncMock(return_value=_fake_completion("not json")),
+        ):
+            with pytest.raises(RuntimeError):
+                await router.route(
+                    messages=[{"role": "user", "content": "x"}],
+                    schema={"type": "object"},
+                )
+
+        # Metrics must still be published even when we raise
+        mock_redis.publish.assert_awaited_once()
+        _, payload_str = mock_redis.publish.call_args[0]
+        payload = json.loads(payload_str)
+        assert payload["model_used"] == "none"