Domain

InteractForge AI

InteractForge turns a single research goal into a complete research package: 1. **Plan** — Gemini Flash creates 5–8 actionable research tasks 2. **Select** — You choose which tasks to investigate 3. **Research** — Deep Research Agent runs with web grounding 4. **Synthesize** — Gemini Pro writes an executive report 5. **Visualize** — Gemini image model creates a TL;DR infographic Each phase chains through `previous_interaction_id`, preserving context across the workflow.

Python Streamlit

← Back to marketplace

InteractForge AI

Gemini Interactions research agent by Sharad Khare.

InteractForge AI is a stateful research planner and executor built on Google's Gemini Interactions API. It plans tasks, runs deep web research, synthesizes executive reports, and generates TL;DR infographics.

Built and maintained by Sharad Khare — AI strategist, full-stack developer, and creator of practical agent workflows.

What this project does

InteractForge turns a single research goal into a complete research package:

1. Plan — Gemini Flash creates 5–8 actionable research tasks

2. Select — You choose which tasks to investigate

3. Research — Deep Research Agent runs with web grounding

4. Synthesize — Gemini Pro writes an executive report

5. Visualize — Gemini image model creates a TL;DR infographic

Each phase chains through previous_interaction_id, preserving context across the workflow.

Why Sharad Khare built InteractForge AI

Research workflows often break context between planning, execution, and reporting. InteractForge demonstrates how Gemini Interactions API enables durable, multi-phase research pipelines in one Streamlit experience.

How it works

Research Goal
   │
   ▼
Phase 1: Plan (gemini-3-flash-preview + google_search)
   │
   ▼
Phase 2: Task Selection + Deep Research Agent (background)
   │
   ▼
Phase 3: Executive Synthesis (gemini-3-pro-preview)
   │
   ▼
Infographic TL;DR (gemini-3-pro-image-preview)

Phase details

| Phase | Model/Agent | Output |

|------|-------------|--------|

| Planning | Gemini 3 Flash | Numbered task plan |

| Deep research | Deep Research Agent | Source-backed findings |

| Synthesis | Gemini 3 Pro | Executive markdown report |

| Infographic | Gemini 3 Pro Image | Visual summary |

Features

Stateful interactions with previous_interaction_id
Task-level selection before deep research
Background research execution with progress polling
Executive report generation
Auto-generated infographic summary
Markdown report download

Quick start

cd interactforge-ai
pip install -r requirements.txt
streamlit run app.py

Required API key

Google API key (Gemini / Google AI Studio)

Example research goals

"Research B2B HR SaaS market in Germany - players, regulations, pricing"
"Analyze opportunities for AI customer support tools in mid-market SaaS"
"Investigate sustainable packaging trends in e-commerce logistics"
"Research fintech compliance requirements for Gen Z banking apps"

Project structure

interactforge-ai/
├── app.py
├── interactforge/
│   ├── config.py
│   └── utils.py
├── requirements.txt
├── pyproject.toml
└── README.md

Topics

Sharad KhareInteractForge AIGemini Interactions APIdeep research agentGoogle Gemini research plannerexecutive report generatorstateful AI conversationsStreamlit research workflowinfographic research summarymulti-phase research automation.

Who this is for

Strategy and market research teams
Consultants producing client-ready research briefs
Founders validating market opportunities
Developers learning Gemini Interactions patterns

Use cases

Build structured research plans from one-line goals
Run selective deep research on high-priority tasks
Generate executive reports with visual TL;DR summaries
Prototype stateful research copilots for internal teams

Author

Website: sharadkhare.in

License

Loading file tree…

Select a file to preview its source.

bash

cd interactforge-ai
pip install -r requirements.txt
streamlit run app.py