feat(bridges): add openrouter + gemini bridges (v2 ladder needed them)

Round 5/6 v2 scripts reference these but the bridges themselves were never
shipped. Now matches existing cerebras/groq/chutes interface (text-prompt
input, env-key auth via ~/.hermes/.env).

- bin/openrouter-bridge.sh: meta-router, free models (llama-3.3-70b/
deepseek-r1/qwen-coder/gemini-flash), pool-key support + auto-fallback
through OPENROUTER_API_KEY/_2/_3 on 401/403/429.
- bin/gemini-bridge.sh: Google AI Studio free tier (gemini-2.5-flash /
2.5-pro / 2.0-flash-exp / 2.5-flash-lite), accepts GEMINI_API_KEY or
GOOGLE_API_KEY.

Trigger: Space rebuild on push to wake stuck status-server.

bin/gemini-bridge.sh

@@ -0,0 +1,75 @@
+#!/usr/bin/env bash
+# Gemini bridge — Google AI Studio free tier (15 RPM, 1M tokens/day on flash).
+# Models: gemini-2.5-pro (paid), gemini-2.5-flash (free), gemini-2.0-flash-exp (free).
+#
+# Usage:
+#   echo "<prompt>" | gemini-bridge.sh [--model fast|pro] [--max-tokens N]
+#   gemini-bridge.sh "prompt as arg"
+set -u
+MODEL="gemini-2.5-flash"
+MAX_TOKENS=2000
+TEMP=0.3
+PROMPT=""
+
+while [[ $# -gt 0 ]]; do
+    case "$1" in
+        --model)
+            case "$2" in
+                fast|small)  MODEL="gemini-2.5-flash" ;;
+                pro|big)     MODEL="gemini-2.5-pro" ;;
+                exp)         MODEL="gemini-2.0-flash-exp" ;;
+                lite)        MODEL="gemini-2.5-flash-lite" ;;
+                *)           MODEL="$2" ;;
+            esac; shift 2 ;;
+        --max-tokens) MAX_TOKENS="$2"; shift 2 ;;
+        --temperature) TEMP="$2"; shift 2 ;;
+        *) PROMPT="$*"; break ;;
+    esac
+done
+[[ -z "$PROMPT" ]] && [[ ! -t 0 ]] && PROMPT=$(cat)
+[[ -z "$PROMPT" ]] && { echo "gemini-bridge: no prompt" >&2; exit 2; }
+
+LOG="$HOME/.surrogate/logs/gemini-bridge.log"
+mkdir -p "$(dirname "$LOG")"
+[[ -f "$HOME/.hermes/.env" ]] && { set -a; source "$HOME/.hermes/.env"; set +a; }
+echo "[$(date '+%H:%M:%S')] model=$MODEL len=${#PROMPT}" >> "$LOG"
+
+KEY="${GEMINI_API_KEY:-${GOOGLE_API_KEY:-}}"
+RESPONSE=$(MODEL="$MODEL" MAX_TOKENS="$MAX_TOKENS" TEMP="$TEMP" KEY="$KEY" \
+python3 -c "
+import json, os, sys, urllib.request, urllib.error
+key = os.environ.get('KEY','')
+if not key:
+    print('gemini-bridge: no GEMINI_API_KEY/GOOGLE_API_KEY', file=sys.stderr); sys.exit(2)
+model = os.environ['MODEL']
+body = {
+    'contents': [{'parts':[{'text': sys.stdin.read()}]}],
+    'generationConfig': {
+        'maxOutputTokens': int(os.environ['MAX_TOKENS']),
+        'temperature': float(os.environ['TEMP']),
+    },
+}
+url = f'https://generativelanguage.googleapis.com/v1beta/models/{model}:generateContent?key={key}'
+req = urllib.request.Request(url,
+    data=json.dumps(body).encode(),
+    headers={'Content-Type':'application/json'})
+try:
+    with urllib.request.urlopen(req, timeout=120) as r:
+        d = json.load(r)
+    cand = d.get('candidates',[{}])[0]
+    parts = cand.get('content',{}).get('parts',[])
+    text = ''.join(p.get('text','') for p in parts)
+    if not text and cand.get('finishReason'):
+        print(f'gemini-bridge: finish_reason={cand[\"finishReason\"]}', file=sys.stderr); sys.exit(1)
+    print(text)
+except urllib.error.HTTPError as e:
+    msg = e.read().decode('utf-8','ignore')[:400]
+    print(f'gemini-bridge HTTP {e.code}: {msg}', file=sys.stderr)
+    sys.exit(1)
+except Exception as e:
+    print(f'gemini-bridge error: {e}', file=sys.stderr); sys.exit(1)
+" <<< "$PROMPT")
+RC=$?
+echo "[$(date '+%H:%M:%S')] rc=$RC bytes=${#RESPONSE}" >> "$LOG"
+[[ $RC -ne 0 ]] && exit $RC
+echo "$RESPONSE"

bin/openrouter-bridge.sh

@@ -0,0 +1,95 @@
+#!/usr/bin/env bash
+# OpenRouter bridge — meta-router across many providers.
+# Free models: qwen/qwen-2.5-coder-32b-instruct:free, deepseek/deepseek-r1:free,
+# meta-llama/llama-3.3-70b-instruct:free, google/gemini-2.0-flash-exp:free.
+#
+# Usage (matches cerebras/groq/chutes interface):
+#   echo "<prompt>" | openrouter-bridge.sh [--model fast|big|free|<id>] [--max-tokens N]
+#   openrouter-bridge.sh "prompt as arg"
+set -u
+MODEL="meta-llama/llama-3.3-70b-instruct:free"
+MAX_TOKENS=2000
+TEMP=0.3
+PROMPT=""
+
+while [[ $# -gt 0 ]]; do
+    case "$1" in
+        --model)
+            case "$2" in
+                fast|small)  MODEL="meta-llama/llama-3.3-70b-instruct:free" ;;
+                big)         MODEL="deepseek/deepseek-r1:free" ;;
+                code|coder)  MODEL="meta-llama/llama-3.3-70b-instruct:free" ;;
+                gemini)      MODEL="google/gemini-2.0-flash-exp:free" ;;
+                free)        MODEL="meta-llama/llama-3.3-70b-instruct:free" ;;
+                *)           MODEL="$2" ;;
+            esac; shift 2 ;;
+        --max-tokens) MAX_TOKENS="$2"; shift 2 ;;
+        --temperature) TEMP="$2"; shift 2 ;;
+        *) PROMPT="$*"; break ;;
+    esac
+done
+[[ -z "$PROMPT" ]] && [[ ! -t 0 ]] && PROMPT=$(cat)
+[[ -z "$PROMPT" ]] && { echo "openrouter-bridge: no prompt" >&2; exit 2; }
+
+LOG="$HOME/.surrogate/logs/openrouter-bridge.log"
+mkdir -p "$(dirname "$LOG")"
+[[ -f "$HOME/.hermes/.env" ]] && { set -a; source "$HOME/.hermes/.env"; set +a; }
+echo "[$(date '+%H:%M:%S')] model=$MODEL len=${#PROMPT}" >> "$LOG"
+
+# Pool support: if OPENROUTER_POOL is set (csv of keys), pick one round-robin.
+# Else try OPENROUTER_API_KEY → OPENROUTER_API_KEY_2 → OPENROUTER_API_KEY_3.
+if [[ -n "${OPENROUTER_POOL:-}" ]]; then
+    IFS=',' read -ra _KEYS <<< "$OPENROUTER_POOL"
+    _N=${#_KEYS[@]}
+    _IDX=$(( ($(date +%s) / 30) % _N ))
+    OPENROUTER_API_KEY="${_KEYS[$_IDX]}"
+fi
+# Auto-fallback: if primary 401s, the python below retries with _2 then _3
+OR_KEYS=""
+for k in OPENROUTER_API_KEY OPENROUTER_API_KEY_2 OPENROUTER_API_KEY_3; do
+    v="${!k:-}"
+    [[ -n "$v" ]] && OR_KEYS="${OR_KEYS}${OR_KEYS:+,}${v}"
+done
+
+RESPONSE=$(MODEL="$MODEL" MAX_TOKENS="$MAX_TOKENS" TEMP="$TEMP" OR_KEYS="$OR_KEYS" \
+python3 -c "
+import json, os, sys, urllib.request, urllib.error
+keys = [k for k in os.environ.get('OR_KEYS','').split(',') if k]
+if not keys:
+    print('openrouter-bridge: no OPENROUTER_API_KEY*', file=sys.stderr); sys.exit(2)
+body = {
+    'model': os.environ['MODEL'],
+    'messages': [{'role':'user','content': sys.stdin.read()}],
+    'max_tokens': int(os.environ['MAX_TOKENS']),
+    'temperature': float(os.environ['TEMP']),
+}
+data = json.dumps(body).encode()
+last_err = ''
+for key in keys:
+    req = urllib.request.Request(
+        'https://openrouter.ai/api/v1/chat/completions',
+        data=data,
+        headers={
+            'Content-Type':'application/json',
+            'Authorization':'Bearer '+key,
+            'HTTP-Referer':'https://axentx.dev/surrogate-1',
+            'X-Title':'Surrogate-1',
+        })
+    try:
+        with urllib.request.urlopen(req, timeout=120) as r:
+            d = json.load(r)
+        print(d.get('choices',[{}])[0].get('message',{}).get('content',''))
+        sys.exit(0)
+    except urllib.error.HTTPError as e:
+        last_err = f'HTTP {e.code}: {e.read().decode(\"utf-8\",\"ignore\")[:300]}'
+        if e.code in (401, 403, 429):
+            continue   # try next key
+        break
+    except Exception as e:
+        last_err = str(e); break
+print(f'openrouter-bridge {last_err}', file=sys.stderr); sys.exit(1)
+" <<< "$PROMPT")
+RC=$?
+echo "[$(date '+%H:%M:%S')] rc=$RC bytes=${#RESPONSE}" >> "$LOG"
+[[ $RC -ne 0 ]] && exit $RC
+echo "$RESPONSE"