refactor: remove airllm references from docs, docstrings, and README (#486)

README.md

@@ -82,7 +82,6 @@ cp .env.example .env
 | `OLLAMA_MODEL` | `qwen3:30b` | Primary model for reasoning and tool calling. Fallback: `llama3.1:8b-instruct` |
 | `DEBUG` | `false` | Enable `/docs` and `/redoc` |
 | `TIMMY_MODEL_BACKEND` | `ollama` | `ollama` \| `airllm` \| `auto` |
-| `AIRLLM_MODEL_SIZE` | `70b` | `8b` \| `70b` \| `405b` |
 | `L402_HMAC_SECRET` | *(default — change in prod)* | HMAC signing key for macaroons |
 | `L402_MACAROON_SECRET` | *(default — change in prod)* | Macaroon secret |
 | `LIGHTNING_BACKEND` | `mock` | `mock` (production-ready) \| `lnd` (scaffolded, not yet functional) |
@@ -177,7 +176,6 @@ timmy chat "Explain self-custody" --backend airllm --model-size 70b
 Or set once in `.env`:
 ```bash
 TIMMY_MODEL_BACKEND=auto
-AIRLLM_MODEL_SIZE=70b
 ```
 
 | Flag  | Parameters  | RAM needed |

TEST_COVERAGE_ANALYSIS.md

@@ -111,7 +111,7 @@ pytest: error: unrecognized arguments: -n --dist worksteal
 ### 4a. Missing Error-Path Testing
 
 Many modules have happy-path tests but lack coverage for:
-- **Graceful degradation paths**: The architecture mandates graceful degradation when Ollama/Redis/AirLLM are unavailable, but most fallback paths are untested (e.g., `cascade.py` lines 563–655)
+- **Graceful degradation paths**: The architecture mandates graceful degradation when Ollama/Redis are unavailable, but most fallback paths are untested (e.g., `cascade.py` lines 563–655)
 - **`brain/client.py`**: Only 14.8% covered — connection failures, retries, and error handling are untested
 - **`infrastructure/error_capture.py`**: 0% — the error capture system itself has no tests
 

docs/AUDIT_REPORT.md

@@ -63,11 +63,11 @@ $ python -m pytest -q
 ## 2. Feature-by-Feature Audit
 
 ### 2.1 Timmy Agent
-**Claimed**: Agno-powered conversational agent backed by Ollama, AirLLM for 70B-405B models, SQLite memory
+**Claimed**: Agno-powered conversational agent backed by Ollama, SQLite memory
 **Verdict: REAL & FUNCTIONAL**
 
 - `src/timmy/agent.py` (79 lines): Creates a genuine `agno.Agent` with Ollama model, SQLite persistence, tools, and system prompt
-- Backend selection (`backends.py`) implements real Ollama/AirLLM switching with Apple Silicon detection
+- Backend selection (`backends.py`) implements real Ollama switching with Apple Silicon detection
 - CLI (`cli.py`) provides working `timmy chat`, `timmy think`, `timmy status` commands
 - Approval workflow (`approvals.py`) implements real human-in-the-loop with SQLite-backed state
 - Briefing system (`briefing.py`) generates real scheduled briefings

docs/REFACTORING_PLAN.md

@@ -100,7 +100,7 @@ Bitcoin Lightning economics. No cloud AI.
   make install && make dev  →  http://localhost:8000
 
 ## What's Here
-  - Timmy Agent (Ollama/AirLLM)
+  - Timmy Agent (Ollama)
   - Mission Control Dashboard (FastAPI + HTMX)
   - Swarm Coordinator (multi-agent auctions)
   - Lightning Payments (L402 gating)

docs/SECURITY.md

@@ -6,7 +6,7 @@ This document outlines the security architecture, threat model, and recent audit
 
 Timmy Time is built on the principle of **AI Sovereignty**. Security is not just about preventing unauthorized access, but about ensuring the user maintains full control over their data and AI models.
 
-1.  **Local-First Execution:** All primary AI inference (Ollama/AirLLM) runs on localhost. No data is sent to third-party cloud providers unless explicitly configured (e.g., Grok).
+1.  **Local-First Execution:** All primary AI inference (Ollama) runs on localhost. No data is sent to third-party cloud providers unless explicitly configured (e.g., Grok).
 2.  **Air-Gapped Ready:** The system is designed to run without an internet connection once dependencies and models are cached.
 3.  **Secret Management:** Secrets are never hard-coded. They are managed via Pydantic-settings from `.env` or environment variables.
 

docs/SOVEREIGN_AGI_RESEARCH.md

@@ -59,7 +59,7 @@ already works.
 | LLM routing | CascadeRouter with circuit breakers | Good |
 | Memory tiers | Hot (MEMORY.md) → Vault (markdown) → Semantic (SQLite+vectors) | Good foundation |
 | Module boundaries | 8 packages with clear responsibilities | Good |
-| Multi-backend LLM | Ollama/AirLLM/Grok/Claude with auto-detection | Good |
+| Multi-backend LLM | Ollama/Grok/Claude with auto-detection | Good |
 | Security posture | CSRF, security headers, secret validation, telemetry off | Good |
 
 ### Architecture Diagram (Current State)
@@ -473,7 +473,7 @@ The proposal enforces a strict 2,000-line limit for `src/timmy/`:
 | `workflow_engine.py` | ~200 | YAML loader, step executor, state machine |
 | `tool_registry.py` | ~200 | Dynamic tool discovery, spawn, health check |
 | `memory_system.py` | ~300 | Hot/Vault/Semantic memory interface (existing) |
-| `backends.py` | ~200 | Ollama/AirLLM/Claude/Grok adapters |
+| `backends.py` | ~200 | Ollama/Claude/Grok adapters |
 | `config.py` | ~150 | Pydantic-settings (existing) |
 | `lightning_wallet.py` | ~200 | L402 handling, invoice generation, balance |
 | `utils/` | ~300 | Shared helpers, logging, serialization |

docs/adr/020-cascade-router-integration.md

@@ -4,7 +4,6 @@
 Proposed
 
 ## Context
-Currently, the Timmy agent (`src/timmy/agent.py`) uses `src/timmy/backends.py` which provides a simple abstraction over Ollama and AirLLM. However, this lacks:
 - Automatic failover between multiple LLM providers
 - Circuit breaker pattern for failing providers
 - Cost and latency tracking per provider
@@ -19,14 +18,13 @@ Integrate the Cascade Router as the primary LLM routing layer for Timmy, replaci
 
 ### Current Flow
 ```
-User Request → Timmy Agent → backends.py → Ollama/AirLLM
+User Request → Timmy Agent → backends.py → Ollama
 ```
 
 ### Proposed Flow
 ```
 User Request → Timmy Agent → Cascade Router → Provider 1 (Ollama)
                                       ↓ (if fail)
-                                 Provider 2 (Local AirLLM)
                                       ↓ (if fail)
                                  Provider 3 (API - optional)
                                       ↓
@@ -41,7 +39,6 @@ User Request → Timmy Agent → Cascade Router → Provider 1 (Ollama)
    - Expose provider status in agent responses
 
 2. **Cascade Router** (`src/router/cascade.py`)
-   - Already supports: Ollama, OpenAI, Anthropic, AirLLM
    - Already has: Circuit breakers, metrics, failover logic
    - Add: Integration with existing `src/timmy/prompts.py`
 
@@ -57,7 +54,6 @@ User Request → Timmy Agent → Cascade Router → Provider 1 (Ollama)
 ### Provider Priority Order
 
 1. **Ollama (local)** - Priority 1, always try first
-2. **AirLLM (local)** - Priority 2, if Ollama unavailable
 3. **API providers** - Priority 3+, only if configured
 
 ### Data Flow

src/timmy/__init__.py

@@ -1 +1 @@
-"""Timmy — Core AI agent (Ollama/AirLLM backends, CLI, prompts)."""
+"""Timmy — Core AI agent (Ollama backend, CLI, prompts)."""

src/timmy/agent.py

@@ -197,113 +197,6 @@ def _resolve_backend(requested: str | None) -> str:
     return "ollama"
 
 
-def _build_tools_list(use_tools: bool, skip_mcp: bool) -> list:
-    """Build the Agno tools list (toolkit + optional MCP servers).
-
-    Args:
-        use_tools: Whether the model supports tool calling.
-        skip_mcp:  If True, omit MCP tool servers.
-
-    Returns:
-        List of Toolkit / MCPTools, possibly empty.
-    """
-    if not use_tools:
-        logger.info("Tools disabled (model too small for reliable tool calling)")
-        return []
-
-    toolkit = create_full_toolkit()
-    tools_list: list = [toolkit]
-
-    # Add MCP tool servers (lazy-connected on first arun()).
-    # Skipped when skip_mcp=True — MCP's stdio transport uses anyio cancel
-    # scopes that conflict with asyncio background task cancellation (#72).
-    if not skip_mcp:
-        try:
-            from timmy.mcp_tools import create_filesystem_mcp_tools, create_gitea_mcp_tools
-
-            gitea_mcp = create_gitea_mcp_tools()
-            if gitea_mcp:
-                tools_list.append(gitea_mcp)
-
-            fs_mcp = create_filesystem_mcp_tools()
-            if fs_mcp:
-                tools_list.append(fs_mcp)
-        except Exception as exc:
-            logger.debug("MCP tools unavailable: %s", exc)
-
-    return tools_list
-
-
-def _build_prompt(use_tools: bool, session_id: str) -> str:
-    """Build the full system prompt with optional memory context.
-
-    Args:
-        use_tools:  Whether tools are enabled (affects prompt tier and context budget).
-        session_id: Session identifier for the prompt.
-
-    Returns:
-        Complete system prompt string.
-    """
-    base_prompt = get_system_prompt(tools_enabled=use_tools, session_id=session_id)
-
-    try:
-        from timmy.memory_system import memory_system
-
-        memory_context = memory_system.get_system_context()
-        if memory_context:
-            # Truncate if too long — smaller budget for small models
-            # since the expanded prompt (roster, guardrails) uses more tokens
-            max_context = 2000 if not use_tools else 8000
-            if len(memory_context) > max_context:
-                memory_context = memory_context[:max_context] + "\n... [truncated]"
-            return (
-                f"{base_prompt}\n\n"
-                f"## GROUNDED CONTEXT (verified sources — cite when using)\n\n"
-                f"{memory_context}"
-            )
-    except Exception as exc:
-        logger.warning("Failed to load memory context: %s", exc)
-
-    return base_prompt
-
-
-def _create_ollama_agent(
-    model_name: str,
-    db_file: str,
-    tools_list: list,
-    full_prompt: str,
-    use_tools: bool,
-) -> Agent:
-    """Construct the Agno Agent with an Ollama model.
-
-    Args:
-        model_name: Resolved Ollama model name.
-        db_file:    SQLite file for conversation memory.
-        tools_list: Pre-built tools list (may be empty).
-        full_prompt: Complete system prompt.
-        use_tools:  Whether tools are enabled.
-
-    Returns:
-        Configured Agno Agent.
-    """
-    model_kwargs = {}
-    if settings.ollama_num_ctx > 0:
-        model_kwargs["options"] = {"num_ctx": settings.ollama_num_ctx}
-
-    return Agent(
-        name="Agent",
-        model=Ollama(id=model_name, host=settings.ollama_url, timeout=300, **model_kwargs),
-        db=SqliteDb(db_file=db_file),
-        description=full_prompt,
-        add_history_to_context=True,
-        num_history_runs=20,
-        markdown=False,
-        tools=tools_list if tools_list else None,
-        tool_call_limit=settings.max_agent_steps if use_tools else None,
-        telemetry=settings.telemetry_enabled,
-    )
-
-
 def create_timmy(
     db_file: str = "timmy.db",
     backend: str | None = None,
@@ -345,12 +238,16 @@ def create_timmy(
         return TimmyAirLLMAgent(model_size=size)
 
     # Default: Ollama via Agno.
+    # Resolve model with automatic pulling and fallback
     model_name, is_fallback = _resolve_model_with_fallback(
         requested_model=None,
         require_vision=False,
         auto_pull=True,
     )
 
+    # If Ollama is completely unreachable, fail loudly.
+    # Sovereignty: never silently send data to a cloud API.
+    # Use --backend claude explicitly if you want cloud inference.
     if not _check_model_available(model_name):
         logger.error(
             "Ollama unreachable and no local models available. "
@@ -361,9 +258,74 @@ def create_timmy(
         logger.info("Using fallback model %s (requested was unavailable)", model_name)
 
     use_tools = _model_supports_tools(model_name)
-    tools_list = _build_tools_list(use_tools, skip_mcp)
-    full_prompt = _build_prompt(use_tools, session_id)
-    agent = _create_ollama_agent(model_name, db_file, tools_list, full_prompt, use_tools)
+
+    # Conditionally include tools — small models get none
+    toolkit = create_full_toolkit() if use_tools else None
+    if not use_tools:
+        logger.info("Tools disabled for model %s (too small for reliable tool calling)", model_name)
+
+    # Build the tools list — Agno accepts a list of Toolkit / MCPTools
+    tools_list: list = []
+    if toolkit:
+        tools_list.append(toolkit)
+
+    # Add MCP tool servers (lazy-connected on first arun()).
+    # Skipped when skip_mcp=True — MCP's stdio transport uses anyio cancel
+    # scopes that conflict with asyncio background task cancellation (#72).
+    if use_tools and not skip_mcp:
+        try:
+            from timmy.mcp_tools import create_filesystem_mcp_tools, create_gitea_mcp_tools
+
+            gitea_mcp = create_gitea_mcp_tools()
+            if gitea_mcp:
+                tools_list.append(gitea_mcp)
+
+            fs_mcp = create_filesystem_mcp_tools()
+            if fs_mcp:
+                tools_list.append(fs_mcp)
+        except Exception as exc:
+            logger.debug("MCP tools unavailable: %s", exc)
+
+    # Select prompt tier based on tool capability
+    base_prompt = get_system_prompt(tools_enabled=use_tools, session_id=session_id)
+
+    # Try to load memory context
+    try:
+        from timmy.memory_system import memory_system
+
+        memory_context = memory_system.get_system_context()
+        if memory_context:
+            # Truncate if too long — smaller budget for small models
+            # since the expanded prompt (roster, guardrails) uses more tokens
+            max_context = 2000 if not use_tools else 8000
+            if len(memory_context) > max_context:
+                memory_context = memory_context[:max_context] + "\n... [truncated]"
+            full_prompt = (
+                f"{base_prompt}\n\n"
+                f"## GROUNDED CONTEXT (verified sources — cite when using)\n\n"
+                f"{memory_context}"
+            )
+        else:
+            full_prompt = base_prompt
+    except Exception as exc:
+        logger.warning("Failed to load memory context: %s", exc)
+        full_prompt = base_prompt
+
+    model_kwargs = {}
+    if settings.ollama_num_ctx > 0:
+        model_kwargs["options"] = {"num_ctx": settings.ollama_num_ctx}
+    agent = Agent(
+        name="Agent",
+        model=Ollama(id=model_name, host=settings.ollama_url, timeout=300, **model_kwargs),
+        db=SqliteDb(db_file=db_file),
+        description=full_prompt,
+        add_history_to_context=True,
+        num_history_runs=20,
+        markdown=False,
+        tools=tools_list if tools_list else None,
+        tool_call_limit=settings.max_agent_steps if use_tools else None,
+        telemetry=settings.telemetry_enabled,
+    )
     _warmup_model(model_name)
     return agent
 

tests/dashboard/test_mobile_scenarios.py

@@ -10,7 +10,7 @@ Categories:
   M3xx  iOS keyboard & zoom prevention
   M4xx  HTMX robustness (double-submit, sync)
   M5xx  Safe-area / notch support
-  M6xx  AirLLM backend interface contract
+  M6xx  Backend interface contract
 """
 
 import re
@@ -208,7 +208,7 @@ def test_M505_dvh_units_used():
     assert "dvh" in css
 
 
-# ── M6xx — AirLLM backend interface contract ──────────────────────────────────
+# ── M6xx — Backend interface contract ──────────────────────────────────
 
 
 def test_M601_airllm_agent_has_run_method():

tests/timmy/test_agent.py

@@ -454,127 +454,3 @@ def test_no_hardcoded_fallback_constants_in_agent():
     assert not hasattr(agent_mod, "VISION_MODEL_FALLBACKS"), (
         "Hardcoded VISION_MODEL_FALLBACKS still exists — use settings.vision_fallback_models"
     )
-
-
-# ── _build_tools_list helper ─────────────────────────────────────────────────
-
-
-def test_build_tools_list_returns_empty_when_no_tools():
-    """When use_tools=False, _build_tools_list returns an empty list."""
-    from timmy.agent import _build_tools_list
-
-    result = _build_tools_list(use_tools=False, skip_mcp=False)
-    assert result == []
-
-
-def test_build_tools_list_includes_toolkit():
-    """When use_tools=True, _build_tools_list includes the toolkit."""
-    mock_toolkit = MagicMock()
-    with patch("timmy.agent.create_full_toolkit", return_value=mock_toolkit):
-        from timmy.agent import _build_tools_list
-
-        result = _build_tools_list(use_tools=True, skip_mcp=True)
-
-    assert mock_toolkit in result
-
-
-def test_build_tools_list_adds_mcp_when_not_skipped():
-    """When skip_mcp=False, _build_tools_list attempts MCP tools."""
-    mock_toolkit = MagicMock()
-    mock_gitea = MagicMock()
-    with (
-        patch("timmy.agent.create_full_toolkit", return_value=mock_toolkit),
-        patch("timmy.mcp_tools.create_gitea_mcp_tools", return_value=mock_gitea),
-        patch("timmy.mcp_tools.create_filesystem_mcp_tools", return_value=None),
-    ):
-        from timmy.agent import _build_tools_list
-
-        result = _build_tools_list(use_tools=True, skip_mcp=False)
-
-    assert mock_toolkit in result
-    assert mock_gitea in result
-
-
-# ── _build_prompt helper ─────────────────────────────────────────────────────
-
-
-def test_build_prompt_returns_base_when_no_memory():
-    """_build_prompt returns base prompt when memory context is empty."""
-    with patch("timmy.memory_system.memory_system") as mock_mem:
-        mock_mem.get_system_context.return_value = ""
-        from timmy.agent import _build_prompt
-
-        result = _build_prompt(use_tools=True, session_id="test")
-
-    assert "Timmy" in result
-
-
-def test_build_prompt_appends_memory_context():
-    """_build_prompt appends memory context when available."""
-    with patch("timmy.memory_system.memory_system") as mock_mem:
-        mock_mem.get_system_context.return_value = "User likes pizza"
-        from timmy.agent import _build_prompt
-
-        result = _build_prompt(use_tools=True, session_id="test")
-
-    assert "User likes pizza" in result
-    assert "GROUNDED CONTEXT" in result
-
-
-def test_build_prompt_truncates_long_memory_for_small_models():
-    """_build_prompt truncates memory for small models (use_tools=False)."""
-    long_context = "x" * 5000
-    with patch("timmy.memory_system.memory_system") as mock_mem:
-        mock_mem.get_system_context.return_value = long_context
-        from timmy.agent import _build_prompt
-
-        result = _build_prompt(use_tools=False, session_id="test")
-
-    # Max context is 2000 for small models + truncation marker
-    assert "[truncated]" in result
-
-
-# ── _create_ollama_agent helper ──────────────────────────────────────────────
-
-
-def test_create_ollama_agent_passes_correct_kwargs():
-    """_create_ollama_agent passes the expected kwargs to Agent()."""
-    with (
-        patch("timmy.agent.Agent") as MockAgent,
-        patch("timmy.agent.Ollama"),
-        patch("timmy.agent.SqliteDb"),
-    ):
-        from timmy.agent import _create_ollama_agent
-
-        _create_ollama_agent(
-            model_name="test-model",
-            db_file="test.db",
-            tools_list=[MagicMock()],
-            full_prompt="Test prompt",
-            use_tools=True,
-        )
-
-        kwargs = MockAgent.call_args.kwargs
-        assert kwargs["description"] == "Test prompt"
-        assert kwargs["tools"] is not None
-
-
-def test_create_ollama_agent_none_tools_when_empty():
-    """_create_ollama_agent passes tools=None when tools_list is empty."""
-    with (
-        patch("timmy.agent.Agent") as MockAgent,
-        patch("timmy.agent.Ollama"),
-        patch("timmy.agent.SqliteDb"),
-    ):
-        from timmy.agent import _create_ollama_agent
-
-        _create_ollama_agent(
-            model_name="test-model",
-            db_file="test.db",
-            tools_list=[],
-            full_prompt="Test prompt",
-            use_tools=False,
-        )
-
-        kwargs = MockAgent.call_args.kwargs
-        assert kwargs["tools"] is None

tests/timmy/test_backends.py

@@ -1,4 +1,4 @@
-"""Tests for src/timmy/backends.py — AirLLM wrapper and helpers."""
+"""Tests for src/timmy/backends.py — backend helpers."""
 
 import sys
 from unittest.mock import MagicMock, patch