r2ai is radare2's language-model plugin; this page configures RedRouter and asks it to explain one function in a lab binary.
Recipe, not run by RedRouter.
Setup
Install radare2 with r2pm and a C compiler. Use an authorized lab binary with a main function. These commands install r2ai and open the binary for analysis.
{
r2pm -Uci r2ai
printf 'RedRouter API key: '
read -r -s REDROUTER_API_KEY
printf '\n'
printf 'Path to your lab binary: '
read -r SAMPLE
OPENAI_API_KEY="$REDROUTER_API_KEY" r2 \
-c 'r2ai -e api=openai' \
-c 'r2ai -e baseurl=https://redrouter.org/v1' \
-c 'r2ai -e model=red-cyber-1' \
-c 'r2ai -e max_tokens=8192' \
-c 'r2ai -e auto.yolo=false' \
-c aa -c 's main' -c 'r2ai -q explain' "$SAMPLE"
}
r2pm -Uci r2ai
$REDROUTER_API_KEY = [System.Net.NetworkCredential]::new('', (Read-Host 'RedRouter API key' -AsSecureString)).Password
$SAMPLE = Read-Host 'Path to your lab binary'
$previousKey = $env:OPENAI_API_KEY
try {
$env:OPENAI_API_KEY = $REDROUTER_API_KEY
r2 -c 'r2ai -e api=openai' `
-c 'r2ai -e baseurl=https://redrouter.org/v1' `
-c 'r2ai -e model=red-cyber-1' `
-c 'r2ai -e max_tokens=8192' `
-c 'r2ai -e auto.yolo=false' `
-c aa -c 's main' -c 'r2ai -q explain' $SAMPLE
} finally {
$env:OPENAI_API_KEY = $previousKey
}
Free key check
GET /v1/key checks your key and its restrictions without generating tokens or spending inference credit. Run it before the analysis request.
{
printf 'RedRouter API key: '
read -r -s REDROUTER_API_KEY
printf '\n'
curl --fail-with-body -sS https://redrouter.org/v1/key \
-H "Authorization: Bearer $REDROUTER_API_KEY"
}
$REDROUTER_API_KEY = [System.Net.NetworkCredential]::new('', (Read-Host 'RedRouter API key' -AsSecureString)).Password
Invoke-RestMethod -Method Get -Uri 'https://redrouter.org/v1/key' `
-Headers @{ Authorization = "Bearer $REDROUTER_API_KEY" }
Known limits
- The source recipe covers r2ai 1.5.4; the package manager may install a newer version.
- Auto mode requires tool calling. Keep
auto.yolo=falseso proposed commands need approval. - Disassembly and strings enter the model context. Send only material you may share and verify claims against radare2's output.
- Large contexts can exceed command-line limits when using curl; r2ai's
http.use_files=truesetting addresses that case.
Read the full r2ai setup guide.