Skip to main content

⚡ Deep Dossier

Documentation GitHub Stars License: MIT

📖 Live Documentation & Interactive FAQs: https://agent-docs-coral.vercel.app

Deep Dossier Live Terminal Demo

An open-source autonomous multi-step research agent with universal model support.

Deep Dossier decomposes complex research prompts into targeted search plans, recursively crawls dozens of web pages and documents in parallel, resolves information gaps, and produces publication-grade Markdown and interactive HTML dossiers with verified in-text citations.

🌐 Universal Model Support (Bring Your Own Key)

Deep Dossier is model-agnostic. Choose your preferred AI provider or run completely local:

Provider Supported Models Environment Variable
Google Gemini gemini-2.5-flash, gemini-2.5-pro, gemini-1.5-pro export GEMINI_API_KEY="..."
Anthropic Claude claude-3-5-sonnet-20241022, claude-3-7-sonnet, claude-3-opus export ANTHROPIC_API_KEY="..."
OpenAI gpt-4o, gpt-4o-mini, o1, o3-mini export OPENAI_API_KEY="..."
DeepSeek deepseek-chat, deepseek-reasoner export DEEPSEEK_API_KEY="..."
NVIDIA Nemotron nvidia/nemotron-4-340b-instruct export NVIDIA_API_KEY="..."
OpenRouter 200+ models across all frontier labs export OPENROUTER_API_KEY="..."
Local / Self-Hosted Ollama (llama3.2, mistral, qwen2.5), vLLM --base-url http://localhost:11434/v1

🌟 Key Highlights

  • Multi-Provider Flexibility: Switch between Gemini, Claude, GPT, DeepSeek, or local Ollama with a single flag.
  • Recursive Search & Discovery: Explores multi-hop questions by identifying knowledge gaps after each crawl iteration.
  • Zero-Config Search: Built-in DuckDuckGo search integration requiring zero additional API keys.
  • Publication Deliverables: Exports structured Markdown, JSON intelligence dossiers, and self-contained dark-mode interactive HTML reports.

🚀 Quick Start

1. Installation

git clone https://github.com/shubhransh-gupta/deep-dossier.git
cd deep-dossier
pip install -e .

2. Set Your Preferred API Key

Choose any provider:

# Option A: Google Gemini
export GEMINI_API_KEY="AIzaSy..."

# Option B: Anthropic Claude
export ANTHROPIC_API_KEY="sk-ant-..."

# Option C: OpenAI
export OPENAI_API_KEY="sk-proj-..."

# Option D: DeepSeek
export DEEPSEEK_API_KEY="sk-..."

3. Run a Research Mission

Using default/auto-detected provider:

deep-dossier "Analyze battery degradation patterns of LFP vs NMC cells in commercial EVs"

Explicitly choose provider and model:

# Using Anthropic Claude 3.5 Sonnet
deep-dossier "Solid-state battery commercial readiness" --provider anthropic --model claude-3-5-sonnet-20241022

# Using OpenAI GPT-4o
deep-dossier "Cross-border payment rail comparison" --provider openai --model gpt-4o

# Using Local Ollama
deep-dossier "Local AI agent architectures" --base-url http://localhost:11434/v1 --model llama3.2

⚙️ Options

deep-dossier [TOPIC] [OPTIONS]

Options:
  --provider, -p TEXT     LLM provider: gemini, openai, anthropic, deepseek, nvidia, ollama, openrouter
  --model, -m TEXT        Specific model name
  --base-url TEXT         Custom API base URL (for Ollama, vLLM, or custom proxies)
  --depth, -d INTEGER     Recursive search depth (default: 2)
  --breadth, -b INTEGER   Number of sources to retrieve per search query (default: 4)
  --output, -o PATH       Output directory for generated dossiers (default: ./research_outputs)
  --api-key, -k TEXT      Explicit API key override
  --no-html               Disable HTML report generation

📦 Project Structure

deep-dossier/
├── deep_dossier/
│   ├── agent.py       # Core autonomous orchestrator and synthesis engine
│   ├── llm.py         # Universal multi-provider LLM interface (Gemini/Claude/GPT/Ollama)
│   ├── crawler.py     # Concurrent HTML & document content extractor
│   ├── search.py      # Search engine interface
│   ├── report.py      # Markdown and interactive HTML renderer
│   ├── config.py      # Configuration and environment loader
│   └── cli.py         # Rich terminal interface
└── pyproject.toml     # Build metadata and entry points

📄 License

MIT License. Copyright (c) 2026 Shubhransh Gupta.

Metadata

Release files for deep-dossier 0.1.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for deep-dossier 0.1.0
File Size Uploaded Provenance Actions
deep_dossier-0.1.0.tar.gz 18.9 kB Without provenance Details for deep_dossier-0.1.0.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for deep-dossier 0.1.0
File Interpreter ABI Platform Actions
deep_dossier-0.1.0-py3-none-any.whl Python 3 none any Details for deep_dossier-0.1.0-py3-none-any.whl

Total release size: 38.5 kB

Release files / deep_dossier-0.1.0.tar.gz

Download URL deep_dossier-0.1.0.tar.gz
Size 18.9 kB
Tags Source
SHA-256 checksum
How to use checksums
3d8f40590bd5501680814c68332ae4dc879da9205e5ce85ceef90e205b5b5139
BLAKE2b-256 checksum
How to use checksums
b7df52271883e5ed5572d8421704c02faba8b236a2fff1ef4d22f53d408a9146
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.14.0

Release files / deep_dossier-0.1.0-py3-none-any.whl

Download URL deep_dossier-0.1.0-py3-none-any.whl
Size 19.6 kB
Tags Python 3
SHA-256 checksum
How to use checksums
e380b88c2514c47b4e648e57674e6939e85652819290ae02594e2ba70b03fcd8
BLAKE2b-256 checksum
How to use checksums
477d29e8962953a943f46a19e3183a242f1732b8473bdc244c874a2c45f2923d
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.14.0

Security for deep-dossier 0.1.0

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Release history Release notifications | RSS feed

This release

0.1.0 This release

2 release files No provenance

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page