TSPilot

KO | EN

gitlite — search

#agent #java #gpt #bash #awesome-list #command-line #javascript #swift #chatgpt #editor

(★ 9)

No description available.

File Explorer

ZIP Download

- __init__.py
- base.py
- data_agent.py
- - __init__.py
  - chat.py
  - resources.py
  - visualizations.py
- __init__.py
- deps.py
- model_config.py
- server.py
- settings.py
- database_catalog.json
- models.json
- timeouts.yaml
- - __init__.py
  - code_policy.py
  - python_runner.py
- - __init__.py
  - catalog.py
  - connection_pool.py
  - connector.py
  - dialects.py
  - engine.py
  - llm_query.py
  - metadata_fetcher.py
  - profile_cache.py
  - query_errors.py
  - result_processor.py
  - schema.py
- - __init__.py
  - action_output.py
  - action_space.py
  - capabilities.py
  - observation.py
  - observation_view.py
  - transition.py
- - __init__.py
  - binder.py
  - contracts.py
  - embedding.py
  - embedding_store.py
  - learning.py
  - memory.py
  - retriever.py
  - runtime.py
- - __init__.py
  - retriever.py
- - composer.py
- - __init__.py
  - registry.py
- - __init__.py
  - anomaly_adapter.py
  - anomaly_registry.py
  - evidence_resolution.py
  - forecast_adapter.py
  - forecast_registry.py
  - normalization.py
- - __init__.py
  - artifact_store.py
  - echarts.py
  - materializer.py
- __init__.py
- artifact_sources.py
- completion.py
- intent.py
- time_range.py
- bitcoin_historical_annotated.csv
- energydata_complete.csv
- - system.md
- - api.md
  - visualization-v5.md
- - database-configuration.md
- README.md
- - App.test.tsx
  - App.tsx
  - databaseCatalog.ts
  - i18n.test.tsx
  - i18n.tsx
  - main.tsx
  - styles.css
  - types.ts
- .gitignore
- index.html
- package-lock.json
- package.json
- tsconfig.json
- tsconfig.node.json
- vite.config.ts
- __init__.py
- data_agent.py
- __init__.py
- action_policy.py
- artifacts.py
- conversation_log.py
- conversation_state.py
- language.py
- llm_trace.py
- output_selection.py
- plain_chat.py
- prompt_locale.py
- react_loop.py
- request_state.py
- timeout_policy.py
- token_usage.py
- tool_executor.py
- trace.py
- __init__.py
- analysis_context.py
- policy.py
- runner.py
- worker.py
- __init__.py
- action_output.py
- agent_turn.py
- analysis.py
- api.py
- database.py
- database_context.py
- echarts_plan.py
- key_insight.py
- output.py
- state.py
- task_contract.py
- timeseries.py
- tool.py
- visual_verification.py
- visualization.py
- backend-dev.sh
- dev.sh
- e2e_echarts_suite.py
- replay_visualization_from_logs.py
- fakes.py
- test_action_policy.py
- test_analysis_artifacts.py
- test_artifact_source_flow.py
- test_chat_api.py
- test_data_agent.py
- test_database_catalog.py
- test_evidence_artifacts.py
- test_format_answer.py
- test_generated_analysis.py
- test_insight_memory_learning.py
- test_insight_memory_persistence.py
- test_insight_memory_settings_api.py
- test_insight_memory_tool_calling.py
- test_insight_observation_visualization.py
- test_intent_profile.py
- test_key_insight_runtime.py
- test_language_policy.py
- test_llm_trace.py
- test_model_config_api.py
- test_multiseries_support.py
- test_prompt_context.py
- test_public_query_visibility.py
- test_sql_query_tool.py
- test_terminate.py
- test_timeout_policy.py
- test_timeseries_model_extensibility.py
- test_todowrite_semantics.py
- test_token_usage.py
- test_tool_evidence_refs.py
- test_visualization_echarts.py
- test_visualization_echarts_replay.py
- __init__.py
- anomaly.py
- base.py
- code_interpreter.py
- forecast.py
- format_answer.py
- rag.py
- registry.py
- skill.py
- sql_query.py
- terminate.py
- todowrite.py
- visualization.py
.gitignore
LICENSE
pyproject.toml
README.md
README_cn.md
requirements.txt
tspilot-page.png
video-en.gif
video-zh.gif

# Use via CDN

jsDelivr

jsDelivr serves any public GitHub repository as a CDN with zero setup. Pick a version and a file to get a ready-to-paste link and snippet.

Version

File

Link

Example

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

<h1 align="center">TSPilot: Agentic AI-native Time Series Data Interaction System</h1>

Connect, query, analyze, and understand time-series data through natural language

Natural-language data interaction&nbsp;&nbsp;·&nbsp;&nbsp;Multiple time-series databases&nbsp;&nbsp;·&nbsp;&nbsp;Full-result analysis 
 Reusable analytical artifacts&nbsp;&nbsp;·&nbsp;&nbsp;Visual verification of analytical conclusions

English&nbsp;&nbsp;|&nbsp;&nbsp;<a href="README_cn.md">简体中文</a>

<a href="#-updates">Updates</a>&nbsp;&nbsp;•&nbsp;&nbsp;
 <a href="#-product-demo">Product demo</a>&nbsp;&nbsp;•&nbsp;&nbsp;
 <a href="#-core-capabilities">Core capabilities</a>&nbsp;&nbsp;•&nbsp;&nbsp;
 <a href="#-quick-start">Quick start</a>&nbsp;&nbsp;•&nbsp;&nbsp;
 <a href="#-database-connectivity">Database connectivity</a>&nbsp;&nbsp;•&nbsp;&nbsp;
 <a href="#-project-documentation">Documentation</a>

---

## 📣 Updates

- **August 23, 2026:** TSPilot v0.1 officially released 🎉🎉.

## 🧭 Why TSPilot

Time-series analysis involves much more than writing a database query. Users often need to understand **data structure, time ranges, aggregation choices, and data quality**, then move repeatedly between querying, analysis, and visualization tools before reaching a reliable conclusion.

TSPilot connects **natural-language interaction** with **time-series querying, analysis, and visualization**. It helps users explore data, detect anomalies, forecast future values, discover patterns, and produce **conclusions grounded in real evidence**.

The artifacts and insights created during one analysis can be **reused in later questions**, allowing an investigation to continue naturally instead of starting from scratch.

## 🎬 Product demo

## ✨ Core capabilities

<table>
 <thead>
 <tr>
 <th width="32%">Capability</th>
 <th>Description</th>
 </tr>
 </thead>
 <tbody>
 <tr><td>💬 Natural-language data interaction</td><td>Query, explore, and analyze time-series data using ordinary language.</td></tr>
 <tr><td>🗄️ Multiple time-series databases</td><td>Connect and manage different time-series databases through one consistent experience.</td></tr>
 <tr><td>📊 Full-result analysis</td><td>Perform statistical analysis, forecasting, anomaly detection, and pattern discovery over complete query results.</td></tr>
 <tr><td>♻️ Reusable analytical artifacts</td><td>Preserve query results, insights, and visualizations for follow-up questions and continued analysis.</td></tr>
 <tr><td>📈 Visual verification</td><td>Combine complete time series, forecast intervals, anomaly points, and key windows in interactive line charts.</td></tr>
 </tbody>
</table>

## 🚀 Quick start

### 1. Prerequisites

- Python 3.10+
- Node.js and npm
- An OpenAI-compatible model endpoint
- An accessible time-series database

### 2. Install

```bash
git clone https://github.com/feilvvl/TSPilot.git
cd TSPilot

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
npm --prefix frontend install
```

On Windows, activate the virtual environment with `.venv\Scripts\activate`.

### 3. Start TSPilot

Open two terminals in the project root and start the backend and frontend separately.

Terminal 1 (backend):

```bash
source .venv/bin/activate
python -m uvicorn app.server:app --host 127.0.0.1 --port 5680
```

Terminal 2 (frontend):

```bash
cd frontend
npm run dev
```

Open:

- Web interface: `http://localhost:5173`
- Backend API: `http://127.0.0.1:5680`
- API documentation: `http://127.0.0.1:5680/docs`

### 4. Complete the initial setup

After startup:

1. Open **Model Management**, add an OpenAI-compatible language model, and provide its API key.
2. Configure an embedding, forecast model, or anomaly detector when needed.
3. Open **Database Management**, add a database, and test the connection.

Model and database settings are stored in the local workspace. No `.env` file is required.

### 5. Ask a question

Select a connected database and ask a question in natural language. TSPilot will organize the required query, calculation, anomaly detection, forecasting, and visualization steps, then preserve the supporting evidence in its final answer.

## 🗄️ Database connectivity

TSPilot provides unified access to 24 time-series database connection types. Databases can be added and tested in the Web interface or maintained as YAML files under `configs/databases/`. Once configured, users can query and analyze data from different sources through a consistent natural-language experience.

<h3 align="center">🔌 Supported databases</h3>

Configure once, then connect, query, and analyze different data sources in a consistent way.

<table align="center">
 <tr>
 <td align="center"><img src="frontend/public/database-logos/influxdb.svg" alt="InfluxDB 2 and InfluxDB 3" height="42"> InfluxDB 2 / 3</td>
 <td align="center"><img src="frontend/public/database-logos/kdb.png" alt="kdb+" height="42"> kdb+</td>
 <td align="center"><img src="frontend/public/database-logos/prometheus.svg" alt="Prometheus" height="42"> Prometheus</td>
 <td align="center"><img src="frontend/public/database-logos/timescaledb.png" alt="TimescaleDB" height="42"> TimescaleDB</td>
 <td align="center"><img src="frontend/public/database-logos/dolphindb.png" alt="DolphinDB" height="42"> DolphinDB</td>
 <td align="center"><img src="frontend/public/database-logos/druid.png" alt="Apache Druid" height="42"> Apache Druid</td>
 </tr>
 <tr>
 <td align="center"><img src="frontend/public/database-logos/questdb.png" alt="QuestDB" height="42"> QuestDB</td>
 <td align="center"><img src="frontend/public/database-logos/tdengine.png" alt="TDengine" height="42"> TDengine</td>
 <td align="center"><img src="frontend/public/database-logos/iotdb.svg" alt="Apache IoTDB" height="42"> Apache IoTDB</td>
 <td align="center"><img src="frontend/public/database-logos/victoriametrics.png" alt="VictoriaMetrics" height="42"> VictoriaMetrics</td>
 <td align="center"><img src="frontend/public/database-logos/griddb.png" alt="GridDB" height="42"> GridDB</td>
 <td align="center"><img src="frontend/public/database-logos/arc.svg" alt="Arc" height="42"> Arc</td>
 </tr>
 <tr>
 <td align="center"><img src="frontend/public/database-logos/m3db.png" alt="M3DB" height="42"> M3DB</td>
 <td align="center"><img src="frontend/public/database-logos/cratedb.png" alt="CrateDB" height="42"> CrateDB</td>
 <td align="center"><img src="frontend/public/database-logos/cnosdb.png" alt="CnosDB" height="42"> CnosDB</td>
 <td align="center"><img src="frontend/public/database-logos/arcadedb.png" alt="ArcadeDB" height="42"> ArcadeDB</td>
 <td align="center"><img src="frontend/public/database-logos/greptimedb.png" alt="GreptimeDB" height="42"> GreptimeDB</td>
 <td align="center"><img src="frontend/public/database-logos/db2.jpg" alt="IBM Db2" height="42"> IBM Db2</td>
 </tr>
 <tr>
 <td align="center"><img src="frontend/public/database-logos/riak-ts.png" alt="Riak TS" height="42"> Riak TS</td>
 <td align="center"><img src="frontend/public/database-logos/bangdb.png" alt="BangDB" height="42"> BangDB</td>
 <td align="center"><img src="frontend/public/database-logos/machbase.png" alt="Machbase Neo" height="42"> Machbase Neo</td>
 <td align="center"><img src="frontend/public/database-logos/openmldb.jpg" alt="OpenMLDB" height="42"> OpenMLDB</td>
 <td align="center"><img src="frontend/public/database-logos/opengemini.png" alt="openGemini" height="42"> openGemini</td>
 <td align="center">＋ More databases Coming soon</td>
 </tr>
</table>

Connection methods and runtime dependencies vary across databases. Refer to the corresponding database documentation for details.

For setup details, see the [database configuration guide](docs/guides/database-configuration.md).

## 📚 Project documentation

| | Goal | Document |
|:---:|---|---|
| 🧭 | Understand system design and capability boundaries | [System architecture](docs/architecture/system.md) |
| 🔗 | Integrate with the chat API | [Chat API contract](docs/contracts/api.md) |
| 📈 | Understand how data and Insights become interactive charts | [Visualization guide](docs/contracts/visualization-v5.md) |
| 🗄️ | Connect a time-series database | [Database configuration guide](docs/guides/database-configuration.md) |

See the [documentation index](docs/README.md) for the complete public documentation set.

## 🤝 Contributing

Issues and pull requests are welcome.

Changes to agent behavior, tool contracts, database connectors, or public APIs should keep a clear scope, include relevant tests, and update the corresponding specification documents.

## 📮 Contact

For questions or collaboration, contact [feilv@hnu.edu.cn](mailto:feilv@hnu.edu.cn).

TSPilot

Owner Repos

File Explorer

# Use via CDN

Was this content helpful?

Recent Feedback

TSPilot

Owner Repos

File Explorer

# Use via CDN

Was this content helpful?

Recent Feedback

Download README