Self-hosted search
that gets smarter.
Search, synthesis, and research for teams that want stronger answers without surrendering their data, workflow, or margin. Deploy privately or use managed delivery, then shape the experience around your product.
from nome_search import NomeSearch
client = NomeSearch(api_key="nsk-...")
# Instant search with synthesis
result = client.search("NVDA earnings Q1 2026", depth="deep")
print(result.answer) # Markdown with inline [1] citations
print(result.citations) # [{url, title, domain}, ...]
# Streaming — aspects arrive as they complete
for event in client.search_stream("Tesla vs BYD EV strategy"):
if event.type == "aspect":
print(f"[{event.name}] {event.content}")
# Autonomous multi-hop research agent
research = client.research(
"How does SpotGamma HIRO indicator work?",
max_iterations=3,
)
print(research.answer) # Full research report
print(research.follow_up_questions)
# Browser interaction
page = client.interact(
url="https://finance.yahoo.com/quote/AAPL",
actions=[
{"action": "scrape", "selector": "table"},
{"action": "screenshot"},
],
)Capabilities
Everything you need to build search-powered products.
Flexible Search Depth
Dial responses for speed, breadth, or deeper investigation depending on the query and the user intent.
Structured Synthesis
Return usable answers instead of a pile of links, with support for timelines, contradictions, key facts, and follow-up work.
Retrieval Cache
Keep repeat work fast and policy-aware while preserving room for fresh results when recency matters.
Browser Workflows
Support modern sites and guided actions when plain search results are not enough for the job.
Extended Research Runs
Handle multi-step investigations that need follow-up questions, synthesis, and durable outputs rather than one-shot lookup.
SSE Streaming
Stream useful progress and partial outputs while longer research completes in the background.
SDKs + MCP Server
Plug the platform into developer tools, internal assistants, and product workflows without rebuilding your own adapter layer.
Private Deployment Options
Keep data residency, governance, and infrastructure control aligned with the environments your team already trusts.
Compounding Quality
Improve the experience over time through review, policy, and product feedback without exposing your internal playbook publicly.
How it works
Delivery Flow
Why Nome
vs. Perplexity, Tavily, Exa, Firecrawl
vs. Perplexity API
Self-hosted. Your data stays yours. Fine-tune on real usage. Price it however you want.
vs. Tavily
Free SearxNG (no per-query API cost). 6-agent synthesis vs. their single-pass. Browser automation included.
vs. Exa
Semantic cache means sub-100ms repeat queries. /interact endpoint for dynamic pages. Full pipeline, not just search.
vs. Firecrawl
Search + synthesis + research agent — not just crawling. Same browser automation plus everything else.
Pricing
Simple, usage-based pricing.
Start free. Scale as you grow. Self-host for unlimited.
Developer
1,000 searches / month
Start free- 1,000 searches / month
- Quick + Standard depths
- 5 req / min rate limit
- Markdown output
- Community support
- Python SDK
Growth
10,000 searches / month
Get Growth- Everything in Developer
- 10,000 searches / month
- All depths (Quick → Ultra)
- SSE streaming
- /research agent endpoint
- 30 req / min
- JSON schema output
- MCP server access
- Priority support
Scale
100,000 searches / month
Get Scale- Everything in Growth
- 100,000 searches / month
- /interact browser automation
- 100 req / min
- Concurrent browser pool
- Custom domain scoring
- Dedicated support
- Usage analytics dashboard
- Webhook notifications
Enterprise
Unlimited searches / month
Contact sales- Everything in Scale
- Unlimited searches
- Self-hosted deployment
- Custom SearxNG instance
- Private model fine-tuning
- SLA + uptime guarantee
- SSO / SCIM (WorkOS)
- Dedicated infrastructure
- White-label API
Feature comparison
| Feature | Developer | Growth | Scale | Enterprise |
|---|---|---|---|---|
| Monthly searches | 1,000 | 10,000 | 100,000 | Unlimited |
| Search depths | Quick, Standard | All (Quick → Ultra) | All + Browser Pool | All + Custom |
| Rate limit | 5 / min | 30 / min | 100 / min | Custom |
| Streaming (SSE) | No | Yes | Yes | Yes |
| /research agent | No | Yes | Yes | Yes |
| /interact browser | No | No | Yes | Yes |
| JSON schema output | No | Yes | Yes | Yes |
| MCP server | No | Yes | Yes | Yes |
| Semantic cache | Shared | Shared | Dedicated | Private |
| Self-hosted option | No | No | No | Yes |
| Overage pricing | Hard block | $0.005 / search | $0.002 / search | Custom |
| Support | Community | Priority | Dedicated | Dedicated + SLA |
Your search engine. Your data. Your margin.
Keep the operating leverage, the policy control, and the user experience. Deploy privately or use managed delivery without handing away your core research system.