You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
-**2026-05-14** 🧠 **Persistent memory CLI**: Added`vibe-trading memory list/show/search/forget` so users can inspect, search, and prune persistent memory without opening `~/.vibe-trading/memory/` by hand. The implementation keeps non-memory CLI startup fast with lazy imports and hardens memory frontmatter scanning by coercing parsed metadata values back to strings before Rich rendering or recall scoring ([#102](https://github.com/HKUDS/Vibe-Trading/pull/102), thanks @Teerapat-Vatpitak).
55
-
-**2026-05-13** 🧭 **Real-price swarm grounding + worker event hygiene**: Swarm runs now pre-fetch market data and ground worker prompts with real prices before analysts start, routing symbol lookups through `_detect_market(code)`and keeping grounding payloads thread-safe ([#93](https://github.com/HKUDS/Vibe-Trading/pull/93)). Worker reports now persist under allowed swarm-run paths, with better `report.md` summary extraction, longer preset timeouts, and redacted/sanitized event previews for cleaner streaming updates ([#84](https://github.com/HKUDS/Vibe-Trading/pull/84)).
56
-
-**2026-05-12** 🧾 **Trust Layer run cards**: Backtests now emit reproducible `run_card.json` and `run_card.md`files alongside artifacts, capturing config and strategy hashes, data sources, scalar metrics, validation payloads, warnings, and artifact checksums. The backtest tool also returns the run-card paths so CLI/API callers can surface the evidence package directly.
47
+
-**2026-05-14** 🧠 Persistent memory is now inspectable from the CLI via`vibe-trading memory list/show/search/forget` ([#102](https://github.com/HKUDS/Vibe-Trading/pull/102), thanks @Teerapat-Vatpitak).
48
+
-**2026-05-13** 🧭 Swarm runs now ground workers with fetched market data and cleaner persisted reports ([#93](https://github.com/HKUDS/Vibe-Trading/pull/93), [#84](https://github.com/HKUDS/Vibe-Trading/pull/84)).
49
+
-**2026-05-12** 🧾 Backtests now emit `run_card.json` and `run_card.md` alongside artifacts for reproducible research runs.
57
50
58
51
<details>
59
52
<summary>Earlier news</summary>
@@ -92,72 +85,127 @@
92
85
93
86
---
94
87
95
-
## 💡 What Is Vibe-Trading?
96
-
97
-
Vibe-Trading is an AI-powered multi-agent finance workspace that turns natural language requests into executable trading strategies, research insights, and portfolio analysis across global markets.
98
-
99
-
### Key Capabilities:
100
-
• **Natural Language → Strategy** — Describe an idea; the agent writes, tests, and exports trading code<br>
101
-
• **6 Data Sources, Zero Config** — A-shares, HK/US, crypto, futures & forex with automatic fallback<br>
• Black-Scholes pricing & full Greeks calculation<br>
153
-
• Technical pattern recognition & detection<br>
154
-
• Portfolio optimization via MVO/Risk Parity/BL
125
+
• Broker-journal behavior diagnostics<br>
126
+
• Rule-based Shadow Account comparisons<br>
127
+
• Exportable audit reports and strategy code
155
128
</div>
156
129
</td>
157
130
</tr>
158
131
</table>
159
132
160
-
## 74 Skills across 8 Categories
133
+
## 💡 What Is Vibe-Trading?
134
+
135
+
Vibe-Trading is an open-source research workspace for turning finance questions into runnable analysis. It connects natural-language prompts to market-data loaders, strategy generation, backtest engines, reports, exports, and persistent research memory.
136
+
137
+
It is designed for research, simulation, and backtesting. It does not execute live trades.
138
+
139
+
---
140
+
141
+
## ✨ What You Can Do
142
+
143
+
| Task | Output |
144
+
|------|--------|
145
+
|**Ask a trading question**| Market research with tools, data, documents, and reusable session context. |
146
+
|**Backtest a strategy idea**| Strategy code, metrics, benchmark context, validation artifacts, and run cards. |
147
+
|**Review your own trades**| Broker-journal parsing, behavior diagnostics, rule extraction, and Shadow Account comparisons. |
148
+
|**Improve repeated research**| Persistent memory and editable skills turn useful routines into reusable workflows. |
149
+
|**Run analyst teams**| Multi-agent research reviews for investment, quant, crypto, macro, and risk workflows. |
150
+
|**Ship usable artifacts**| Reports, TradingView Pine Script, TDX, MetaTrader 5, MCP tools, and later research sessions. |
151
+
152
+
---
153
+
154
+
## ⚡ Quick Example
155
+
156
+
```bash
157
+
pip install vibe-trading-ai
158
+
vibe-trading run -p "Backtest a BTC-USDT 20/50 moving-average strategy for 2024, summarize return and drawdown, then export the report"
159
+
```
160
+
161
+
```bash
162
+
vibe-trading --upload trades_export.csv
163
+
vibe-trading run -p "Analyze my trading behavior, extract my shadow strategy, and compare it with my actual trades"
164
+
```
165
+
166
+
---
167
+
168
+
## 👥 Shadow Account
169
+
170
+
Shadow Account starts from your own trading records instead of a generic strategy template.
171
+
172
+
Upload a broker export, let the agent summarize your behavior, then compare the actual trading path with a rule-based shadow strategy.
173
+
174
+
| Step | Agent output |
175
+
|------|--------------|
176
+
|**1. Read your journal**| Parses broker exports from 同花顺, 东方财富, 富途, and generic CSV formats. |
177
+
|**2. Profile your behavior**| Holding days, win rate, PnL ratio, drawdown, disposition effect, overtrading, momentum chasing, and anchoring checks. |
178
+
|**3. Extract your rules**| Turns recurring entries/exits into an explicit strategy profile instead of a hand-wavy summary. |
179
+
|**4. Run the shadow**| Backtests the extracted rules and highlights rule breaks, early exits, missed signals, and alternative trade paths. |
180
+
|**5. Deliver the report**| Produces an HTML/PDF report that can be inspected, archived, or refined in a later session. |
181
+
182
+
```bash
183
+
vibe-trading --upload trades_export.csv
184
+
vibe-trading run -p "Analyze my trading behavior, extract my shadow strategy, and compare it with my actual trades"
185
+
```
186
+
187
+
---
188
+
189
+
## 🧪 Research Workflow
190
+
191
+
Most runs follow the same evidence path: route the request, load the right market context, execute tools, validate outputs, and keep the artifacts inspectable.
192
+
193
+
| Layer | What happens |
194
+
|-------|--------------|
195
+
|**Plan**| Selects the relevant finance skills, tools, data sources, and swarm preset when useful. |
196
+
|**Ground**| Pulls A-shares, HK/US equities, crypto, futures, forex, documents, or web context through the available loaders. |
197
+
|**Execute**| Generates testable strategy code, runs tools, and uses the matching backtest engine or analysis workflow. |
198
+
|**Validate**| Adds metrics, benchmark comparison, Monte Carlo, Bootstrap, Walk-Forward, run cards, and warnings where applicable. |
199
+
|**Deliver**| Returns reports, artifacts, tool traces, and exports for TradingView, TDX, MetaTrader 5, MCP clients, or later sessions. |
200
+
201
+
---
202
+
203
+
## 🔩 Detailed Capabilities
204
+
205
+
Detailed inventories are folded below to keep the main README scannable. Open them when you want to inspect the available building blocks.
206
+
207
+
<details>
208
+
<summary><b>Finance Skill Library</b> <sub>74 skills across 8 categories</sub></summary>
161
209
162
210
- 📊 74 specialized finance skills organized into 8 categories
163
211
- 🌐 Complete coverage from traditional markets to crypto & DeFi
@@ -174,7 +222,10 @@ Vibe-Trading is an AI-powered multi-agent finance workspace that turns natural l
|`<PROVIDER>_API_KEY`| Yes*| API key (`OPENROUTER_API_KEY`, `DEEPSEEK_API_KEY`, etc.) |
340
400
|`<PROVIDER>_BASE_URL`| Yes | API endpoint URL |
341
-
|`LANGCHAIN_MODEL_NAME`| Yes | Model name (e.g. `deepseek/deepseek-v3.2`) |
401
+
|`LANGCHAIN_MODEL_NAME`| Yes | Model name (e.g. `deepseek-v4-pro`) |
342
402
|`TUSHARE_TOKEN`| No | Tushare Pro token for A-share data (falls back to AKShare) |
343
403
|`TIMEOUT_SECONDS`| No | LLM call timeout, default 120s |
344
404
|`API_AUTH_KEY`| Recommended for network deployments | Bearer token required when the API is reachable from non-local clients |
@@ -357,10 +417,10 @@ Vibe-Trading is a tool-heavy agent — skills, backtests, memory, and swarms all
357
417
| Tier | Examples | When to use |
358
418
|------|----------|-------------|
359
419
|**Best**|`anthropic/claude-opus-4.7`, `anthropic/claude-sonnet-4.6`, `openai/gpt-5.4`, `google/gemini-3.1-pro-preview`| Complex swarms (3+ agents), long research sessions, paper-grade analysis |
360
-
|**Sweet spot** (default) |`deepseek/deepseek-v3.2`, `x-ai/grok-4.20`, `z-ai/glm-5.1`, `moonshotai/kimi-k2.5`, `qwen/qwen3-max-thinking`| Daily driver — reliable tool-calling at ~1/10 the cost |
420
+
|**Sweet spot** (default) |`deepseek-v4-pro`, `deepseek/deepseek-v4-pro`, `x-ai/grok-4.20`, `z-ai/glm-5.1`, `moonshotai/kimi-k2.5`, `qwen/qwen3-max-thinking`| Daily driver — reliable tool-calling at ~1/10 the cost |
361
421
|**Avoid for agent use**|`*-nano`, `*-flash-lite`, `*-coder-next`, small / distilled variants | Tool-calling is unreliable — the agent will appear to "answer from memory" instead of loading skills or running backtests |
362
422
363
-
The default `agent/.env.example` ships with `deepseek/deepseek-v3.2` — the cheapest option in the sweet-spot tier.
423
+
The default `agent/.env.example` ships with DeepSeek official API + `deepseek-v4-pro`; OpenRouter users can use `deepseek/deepseek-v4-pro`.
0 commit comments