Reverse-engineering research atlas
The recoverable API surface of LinkedIn's first iOS clients (2008–2010)
Static analysis of five archived, decrypted builds of com.linkedin.LinkedIn (1.0 → 3.0.1) using REA and Ghidra 12.1.4. Every claim traces to a byte offset in a specific binary. Historical routes are research records — they are not proxied and no LinkedIn data is served.
Overview
Metrics are computed from stored evidence. Zero verified live endpoints is a valid, honest result.
Most strongly evidenced endpoints
Ranked by static confidence. Click through for the exact evidence trail.
| Method | Historical path | Feature | Confidence | Versions |
|---|---|---|---|---|
| GET | /v2/messages | inbox | confirmed | 1.5, 3.0.1 |
| GET | /v2/messages/count.xml | inbox | confirmed | 1.5 |
| POST | /oauth/login_authorize | authentication | confirmed | 1.0, 1.0.1, 1.0.4, 1.5, 3.0.1 |
| POST | /oauth/request_token | authentication | confirmed | 1.0, 1.0.1, 1.0.4, 1.5, 3.0.1 |
| GET | /v2/members/{memberId}/connections?fields=... | connections | strong | 1.5, 3.0.1 |
| GET | /profile | profile | strong | 1.0, 1.0.1, 1.0.4, 1.5, 3.0.1 |
By confidence
What this atlas is (and is not)
Clarity about provenance is a hard requirement of this project.
Historical endpoints
Recovered from binary evidence. Shown with method, path template, parameters and confidence. These are records, not live services.
This atlas's own API
A first-party read-only JSON API over the evidence. See API docs →
Lawfully verified externals
Only if a permitted route returns real structured data. Currently: 0. Zero is an honest result.