alpaca_live

Trading arena fed by Alpaca data via trtools2. MAF never calls Alpaca directly — trtools2 owns the Alpaca ingestion (live feed engine writes bars/news to Redis Streams; the dashboard's HTTP API serves historical queries from QuestDB). Two routes into the same data: • Hot path — trtools2_bars / trtools2_news Redis Streams. Sub-second latency, populated by the live feed engine. • Warm path — trtools2_api HTTP client. Used when the analyst needs bars over a specific window or wants the per-symbol coverage stats. Bring up requires: trtools2 feed engine running, trtools2 dashboard on :8888. The Setup tab's freshness badges flag whether each route is live.

manual

Phase Pipeline

analysis

parallel

trtools2_analyst (7 tools)

→

synthesis

sequential

synthesis_agent

Ticker

Date

Configure the arena's metadata, data sources, agents and phases. Changes write to the YAML file. Restart the app to apply.

Metadata

Description One-line summary shown on the dashboard. Schedule Cron expression. Leave blank for manual-only runs. Max iterations How many replan loops are allowed before giving up. Target key Drives which outbox publishes the verdict.

Data sources

Bind data adapters to symbolic names. Agents reference these names in their sources list.

Agents

Each agent runs a ReAct loop, calls its bound sources as tools, and emits a structured signal.

Phases

Phases run in order. parallel fans out specialists, sequential runs one-by-one, debate does N-way rounds + judge.

Smart triggers

Auto-run this arena when a Redis Stream event matches a rule. Rules run through safe_eval (boolean/comparisons, abs(), payload.field). Reference →

Ticker

Start Date

End Date

Edit arena configuration (JSON). Changes write to YAML file — restart to apply.

{
  "applicable_target_types": [],
  "description": "Trading arena fed by Alpaca data via trtools2. MAF never calls Alpaca\ndirectly \u2014 trtools2 owns the Alpaca ingestion (live feed engine writes\nbars/news to Redis Streams; the dashboard\u0027s HTTP API serves historical\nqueries from QuestDB).\n\nTwo routes into the same data:\n  \u2022 Hot path \u2014 trtools2_bars / trtools2_news Redis Streams. Sub-second\n    latency, populated by the live feed engine.\n  \u2022 Warm path \u2014 trtools2_api HTTP client. Used when the analyst needs\n    bars over a specific window or wants the per-symbol coverage stats.\n\nBring up requires: trtools2 feed engine running, trtools2 dashboard on\n:8888. The Setup tab\u0027s freshness badges flag whether each route is live.\n",
  "llm": {
    "deep_model": "auto",
    "deep_provider": "ollama",
    "default_provider": "ollama",
    "providers": {
      "ollama": {
        "api_key": "529224e623bb4df6bd2d35ca70830b11.1OcHilLKpOBPZbD1VizrA21P",
        "base_url": "https://ollama.com/v1",
        "max_output_tokens": 4096,
        "model": "auto",
        "temperature": 0.3
      }
    },
    "quick_model": "auto",
    "quick_provider": "ollama"
  },
  "max_iterations": 1,
  "memory": {
    "chromadb_path": "./data/chromadb",
    "embedding_model": "text-embedding-004",
    "embedding_provider": "gemini",
    "instances": []
  },
  "name": "alpaca_live",
  "output_stream": "maf:arena:alpaca_live:output",
  "phases": [
    {
      "agents": [
        {
          "applicable_target_types": [],
          "extra": {},
          "llm_tier": "quick",
          "max_react_steps": 6,
          "name": "trtools2_analyst",
          "role": "specialist",
          "sources": [
            "bars_1m",
            "bars_1h",
            "news_live",
            "strategy_events",
            "history_bars",
            "snapshot",
            "feed_health"
          ],
          "system_prompt": "You are a trading analyst with Alpaca data sourced through\ntrtools2. You see live 1m + 1h bars, news with sentiment,\nstrategy-engine signals, historical daily bars over the last\n~10 weeks, and a snapshot of the latest trade.\n\nFirst check the feed_stats \u2014 if the live feed is stale, mark\nthe data quality as degraded and lower your confidence.\n\nOutput a single AgentSignal JSON with:\n  signal: BULLISH | BEARISH | NEUTRAL\n  confidence: 0.0-1.0\n  summary: one sentence on the dominant driver\n  key_factors: 3-5 short bullet strings (each grounded in one of the data sources)\n",
          "tools": []
        }
      ],
      "max_rounds": 1,
      "name": "analysis",
      "pattern": "parallel",
      "signal_extract": false,
      "transition": "synthesis"
    },
    {
      "agents": [
        {
          "applicable_target_types": [],
          "extra": {},
          "llm_tier": "deep",
          "max_react_steps": 1,
          "name": "synthesis_agent",
          "role": "synthesis",
          "sources": [],
          "system_prompt": "Synthesise the analyst\u0027s signal into a final verdict.\nVocabulary: BUY | HOLD | SELL.\n\nRules of thumb:\n  \u2022 HOLD when the live feed is stale or data sources are mostly empty.\n  \u2022 HOLD when the analyst\u0027s confidence is below 0.55.\n  \u2022 Otherwise propagate the analyst\u0027s direction with a slightly damped confidence.\n",
          "tools": []
        }
      ],
      "max_rounds": 1,
      "name": "synthesis",
      "pattern": "sequential",
      "signal_extract": false,
      "transition": "END"
    }
  ],
  "sources": [
    {
      "adapter": "trtools2_bars",
      "applicable_target_types": [],
      "config": {
        "limit": 100,
        "timeframe": "1m"
      },
      "name": "bars_1m"
    },
    {
      "adapter": "trtools2_bars",
      "applicable_target_types": [],
      "config": {
        "limit": 100,
        "timeframe": "1h"
      },
      "name": "bars_1h"
    },
    {
      "adapter": "trtools2_news",
      "applicable_target_types": [],
      "config": {
        "limit": 25
      },
      "name": "news_live"
    },
    {
      "adapter": "trtools2_strategy_events",
      "applicable_target_types": [],
      "config": {
        "limit": 50
      },
      "name": "strategy_events"
    },
    {
      "adapter": "trtools2_api",
      "applicable_target_types": [],
      "config": {
        "limit": 50,
        "query_type": "bars",
        "timeframe": "1d"
      },
      "name": "history_bars"
    },
    {
      "adapter": "trtools2_api",
      "applicable_target_types": [],
      "config": {
        "query_type": "snapshot"
      },
      "name": "snapshot"
    },
    {
      "adapter": "trtools2_api",
      "applicable_target_types": [],
      "config": {
        "query_type": "feed_stats"
      },
      "name": "feed_health"
    }
  ],
  "target_key": "ticker",
  "triggers": [
    {
      "action_mode": "manual",
      "cooldown_s": 120,
      "name": "kronos high confidence on watched ticker",
      "on_stream": "kronos:forecasts:emitted",
      "target": {
        "ticker": "{payload.symbol}"
      },
      "when": "payload.prob_up \u003e 0.75 or payload.prob_up \u003c 0.25"
    },
    {
      "action_mode": "semi",
      "cooldown_s": 60,
      "name": "trtools2 strategy event",
      "on_stream": "trtools2:strategy:events",
      "target": {
        "ticker": "{payload.symbol}"
      },
      "when": "payload.action == \u0027BUY\u0027 or payload.action == \u0027SELL\u0027"
    }
  ]
}

Data sources bound to this arena. Agents access these via ReAct tool calls.

Source Name	Adapter	Module	Parameters	Used By
bars_1m	trtools2_bars	trtools2	—	trtools2_analyst
bars_1h	trtools2_bars	trtools2	—	trtools2_analyst
news_live	trtools2_news	trtools2	—	trtools2_analyst
strategy_events	trtools2_strategy_events	trtools2	—	trtools2_analyst
history_bars	trtools2_api	trtools2	—	trtools2_analyst
snapshot	trtools2_api	trtools2	—	trtools2_analyst
feed_health	trtools2_api	trtools2	—	trtools2_analyst

Decision trail — every arena run is saved for review and auditing.

Loading memory data...

alpaca_live

Phase Pipeline

analysis

synthesis

Agent Signals

Execution Trace

Source Monitoring

Agent Reports

Debates

Decisions

Target Context

Metadata

Data sources

Agents

Phases

Smart triggers

Performance Metrics

vs Buy & Hold

Equity Curve

Trades

Signal Timeline