Simulate how different creators, platforms, or budget splits would perform before spending money
Test how a mid-campaign change would affect the following two weeks
Attribute past campaign performance by replaying it with different choices
Explore the causal logic using the included 21,000-item demo dataset
| oranai-ltd/oransim | gudong2003/xianyu-auto-reply-fix | lyra81604/zhengxi-views | |
|---|---|---|---|
| Stars | 1,102 | 1,084 | 1,151 |
| Language | Python | Python | Python |
| Last pushed | — | — | 2026-06-30 |
| Maintenance | — | — | Active |
| Setup difficulty | moderate | moderate | moderate |
| Complexity | 3/5 | 3/5 | 3/5 |
| Audience | pm founder | developer | researcher |
Figures from each repo's GitHub metadata at analysis time.
The open-source version ships with a 21,000-item demo dataset, a paid license is needed for the larger real-world dataset.
Oransim is a marketing simulation tool aimed at enterprise marketing teams, specifically the people who decide where to spend campaign budgets. The core problem it addresses is that traditional marketing analysis tools tell you what already happened, but can't answer "what would have happened if we'd made a different decision?", for example, if you'd picked different content creators, shifted budget to a different platform, or changed your timing mid-campaign. Oransim uses a technique called causal simulation: it models a large virtual population of consumers and runs "what if" experiments (called counterfactuals) against that population. Instead of A/B testing a campaign for two weeks at real cost, you can simulate it in about 60 seconds. The engine can rank combinations of creative content, creator choices, and budget allocations before you spend any money, show you how a mid-campaign change (like swapping one creator for another on day three) would affect outcomes over the following two weeks, and attribute performance after a campaign ends by replaying it with different platform choices. The open-source version runs on a 21,000-item demo dataset so you can inspect the full causal logic yourself. An enterprise data license unlocks a much larger dataset of indexed social media content and creator profiles. The tool is built in Python and the causal engine is open source under the Apache-2.0 license. The full README is longer than what was provided.
A causal simulation tool that lets marketing teams test what if campaign decisions in about 60 seconds instead of running real A/B tests.
Mainly Python. The stack also includes Python.
The open-source causal engine can be used freely, including commercially, under the Apache-2.0 license.
Setup difficulty is rated moderate, with roughly 1h+ to a first successful run.
Mainly pm founder.
This repo across BitVibe Labs
Verify against the repo before relying on details.