# Managing Typesense

> Use when working with Typesense — typesense collection management, search tuning, schema analysis, and cluster health monitoring. Covers collection schemas, search analytics, synonym management, override rules, API key scoping, and curation configuration. Read this skill before any Typesense operations.

- Skill: `cloudthinker-ai/managing-typesense` (Agent Skill)
- Install (CLI): `npx skillmds@latest add cloudthinker-ai/managing-typesense`
- Raw SKILL.md: https://api.skillmd.com/api/skills/cloudthinker-ai/managing-typesense/raw
- Safety review: pending (external: skill-scanner PASS, skillspector CAUTION)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Data & Analytics
- Author: cloudthinker-ai (https://skillmd.com/u/cloudthinker-ai)
- Updated: 2026-09-08
- Page: https://skillmd.com/skills/cloudthinker-ai/managing-typesense

---


# Typesense Management Skill

Monitor, analyze, and optimize Typesense search clusters safely.

## MANDATORY: Discovery-First Pattern

**Always check cluster health and list collections before any search or schema operations. Never assume collection names or field types.**

### Phase 1: Discovery

```bash
#!/bin/bash

TS_URL="${TYPESENSE_URL:-http://localhost:8108}"
TS_KEY="${TYPESENSE_API_KEY}"

ts_get() {
    curl -s -H "X-TYPESENSE-API-KEY: $TS_KEY" "$TS_URL$1"
}

echo "=== Cluster Health ==="
ts_get "/health"

echo ""
echo "=== Cluster Metrics ==="
ts_get "/metrics.json" | python3 -c "
import sys, json
data = json.load(sys.stdin)
print(f\"Memory: {data.get('system_memory_used_bytes',0)//1048576}MB / {data.get('system_memory_total_bytes',0)//1048576}MB\")
print(f\"Disk: {data.get('system_disk_used_bytes',0)//1073741824}GB / {data.get('system_disk_total_bytes',0)//1073741824}GB\")
print(f\"Typesense memory: {data.get('typesense_memory_active_bytes',0)//1048576}MB\")
" 2>/dev/null

echo ""
echo "=== Collections ==="
ts_get "/collections" | python3 -c "
import sys, json
for c in json.load(sys.stdin):
    print(f\"Collection: {c['name']} | Docs: {c.get('num_documents',0)} | Fields: {len(c.get('fields',[]))}\")
" 2>/dev/null
```

**Phase 1 outputs:** Cluster health, memory/disk usage, collection list with doc counts and field counts.

### Phase 2: Analysis

```bash
#!/bin/bash

TS_URL="${TYPESENSE_URL:-http://localhost:8108}"
TS_KEY="${TYPESENSE_API_KEY}"

ts_get() {
    curl -s -H "X-TYPESENSE-API-KEY: $TS_KEY" "$TS_URL$1"
}

COLLECTION="${1:-my_collection}"

echo "=== Collection Schema: $COLLECTION ==="
ts_get "/collections/$COLLECTION" | python3 -c "
import sys, json
c = json.load(sys.stdin)
print(f\"Name: {c['name']} | Docs: {c.get('num_documents',0)} | Default sorting: {c.get('default_sorting_field','none')}\")
for f in c.get('fields', []):
    print(f\"  {f['name']}: {f['type']} | facet={f.get('facet',False)} index={f.get('index',True)} optional={f.get('optional',False)}\")
" 2>/dev/null

echo ""
echo "=== Synonyms ==="
ts_get "/collections/$COLLECTION/synonyms" | python3 -c "
import sys, json
data = json.load(sys.stdin)
for s in data.get('synonyms', []):
    print(f\"  {s.get('id','?')}: {s.get('synonyms', s.get('root','?'))}\")
" 2>/dev/null || echo "No synonyms configured"

echo ""
echo "=== Overrides ==="
ts_get "/collections/$COLLECTION/overrides" | python3 -c "
import sys, json
data = json.load(sys.stdin)
for o in data.get('overrides', []):
    print(f\"  {o.get('id','?')}: rule={o.get('rule',{})}\")
" 2>/dev/null || echo "No overrides configured"
```

## Output Format

```
TYPESENSE ANALYSIS
==================
Health: [ok/unhealthy] | Memory: [used]/[total]
Collections: [count] | Total Docs: [count]

ISSUES FOUND:
- [issue with affected collection]

RECOMMENDATIONS:
- [actionable recommendation]
```

## Anti-Hallucination Rules

1. **NEVER assume resource names** — always discover via CLI/API in Phase 1 before referencing in Phase 2.
2. **NEVER fabricate metric names or dimensions** — verify against the service documentation or `--help` output.
3. **NEVER mix CLI commands between service versions** — confirm which version/API you are targeting.
4. **ALWAYS use the discovery → verify → analyze chain** — every resource referenced must have been discovered first.
5. **ALWAYS handle empty results gracefully** — an empty response is valid data, not an error to retry.

## Counter-Rationalizations

| Shortcut | Counter | Why |
|----------|---------|-----|
| "I'll skip discovery and check known resources" | Always run Phase 1 discovery first | Resource names change, new resources appear — assumed names cause errors |
| "The user only asked for a quick check" | Follow the full discovery → analysis flow | Quick checks miss critical issues; structured analysis catches silent failures |
| "Default configuration is probably fine" | Audit configuration explicitly | Defaults often leave logging, security, and optimization features disabled |
| "Metrics aren't needed for this" | Always check relevant metrics when available | API/CLI responses show current state; metrics reveal trends and intermittent issues |
| "I don't have access to that" | Try the command and report the actual error | Assumed permission failures prevent useful investigation; actual errors are informative |