Osintgram is a GPL-3.0 licensed OSINT tool for Instagram, written in Python 3.10+, that gives security researchers, penetration testers and investigators an interactive interface for profiling any public Instagram account from its nickname alone.

What it is

Osintgram is an open-source information-gathering tool that lives in the Python Instagram API ecosystem and runs entirely on the analyst's own machine. Point it at an Instagram account and it collects, cross-references and lays out everything a public profile gives away: followers, captions, hashtags, geotagged locations, posting habits and public contact details. It is driven two ways. AI mode takes a plain-language question and a local model, served through Ollama, chooses which lookups to run and summarizes the results. Base mode needs no model at all: all 28 commands are listed one line each, grouped and searchable, and the analyst ticks what to run, sets parameters and executes them in order. A username is optional in AI mode, so hashtag and place searches can be run against no account at all.

The concrete problem it solves is the reconnaissance workflow that returns a wall of raw JSON from paid third-party lookups with no warning about what a search will cost. Osintgram prices a selection up front, shows remaining credit, caches every request so the same lookup is never paid for twice, and lets a run be capped or stopped server-side rather than only in the browser. It replaces that raw output with presentation: post thumbnails in a filterable grid, geotagged locations on a map, a weekday by hour heatmap of posting activity, and side-by-side comparison of two accounts — with the underlying JSON still one click away on every card. A finished search can be turned into a standalone HTML report that opens in any browser with no server and no Osintgram installed.

Key capabilities

• Two operating modes: AI mode using a local Ollama model that can only choose from the same command list and cannot invent lookups or data, and Base mode requiring no model.
• Profile lookups returning bio and every link in it, follower, following and post counts, category, public email, phone and business address down to coordinates, a linked Facebook id, and Instagram's own "About this account" data: country of registration, creation date and how many times the username changed.
• Network analysis covering followers and followings, Instagram-suggested related accounts, who tagged the target and who the target tagged, most frequent commenters, and mutual connections between two accounts.
• Content analysis including hashtags ranked by use, captions, every comment, like and comment statistics, a photo, video and carousel breakdown with views and paid partnerships, a posting-times heatmap, geotagged locations on a map and photo alt text.
• Target-free searches for posts published with a hashtag, top or recent, or from a place, along with who published them.
• Media handling where stories, highlights and profile pictures come back as direct CDN links, and any card containing media offers a one-click zip of the originals.
• A local web interface served from 127.0.0.1, so the AI runs locally with no OpenAI key, no subscription and no prompt sent to a third party.

Who uses it and how

• Penetration testers and red teams performing information gathering during an authorized engagement against a named Instagram account.
• OSINT analysts and investigators building a picture of a public profile through mutual connections, commenting patterns and account-comparison views.
• Analysts who must hand results to someone else: one button turns a finished search into a standalone HTML report that needs no server and no Osintgram to open.
• Privacy-conscious operators who cannot send prompts or queries to a hosted AI service, since inference runs through Ollama locally and the interface binds to 127.0.0.1.
• Researchers working on a topic rather than a person, leaving the username empty in AI mode and querying a hashtag or a place instead.

Getting started

Install the dependencies with pip install -r requirements.txt, then start the interface with uvicorn src.web.app:app --host 127.0.0.1 --port 8000 --reload, or simply make run. A data backend is required: the page opens on a panel where a HikerAPI key can be pasted, and the key is verified before it is accepted and can be saved for next time.

How it compares

The facts provided name no comparable OSINT tools, only a data backend and a supporting library, so Osintgram stands alone in this registry. Any comparison against commercial profiling services would rest on licence, self-hosting and data ownership, and none of those products are described in the available material.

When to use it — and when not to

Osintgram is the right choice for an analyst who wants local inference, visible lookup costs and a shareable report rather than a hosted dashboard. A self-hoster must still operate a credit-based data backend through a HikerAPI key, because Instagram data is not free, and AI mode additionally requires a local Ollama model to be installed and running. Anyone who wants a hosted service with no local setup, or a library purely for scripted command-line automation, should look elsewhere; the vendored README material is also partly truncated, so backend setup details should be confirmed against the repository.

pip install -r requirements.txt
uvicorn src.web.app:app --host 127.0.0.1 --port 8000 --reload   # or: make run

ollama pull llama3.1:8b     # ~4.7 GB, the default

docker compose up --build      # or: make docker

pip install -r requirements-dev.txt
python -m pytest