SurfSense
Open-source web research infrastructure for AI agents, with self-hosted and usage-based cloud deployment options.
About SurfSense
SurfSense gives AI agents live, structured web data through connectors, a REST API, and an MCP server. It gathers material from Reddit, YouTube, TikTok, Amazon, Google Maps, Google Search, and ordinary web pages for agents, automations, briefs, alerts, and monitoring. The full platform can run on a team’s own infrastructure with its own model-provider keys. A hosted service is also available without a recurring subscription. Docker is the quickest route to a self-hosted installation; manual setup gives teams control over individual components. Hosted usage comes from a prepaid balance and is charged per returned item or crawled page. Failed calls are not billed. Completed results can be written to Notion, Slack, Linear, and Jira. Enterprise arrangements can add custom connectors, on-premises or VPC deployment, audit logs, OIDC or SAML SSO, managed updates, SLA commitments, and dedicated support. Self-hosting suits teams that want infrastructure and billing control and are prepared to handle operations. The cloud avoids that setup, though its cost changes with workload.
Best for
Best use cases
A scheduled agent gathers current material from several web sources and produces a recurring market brief.
An application calls connector APIs for structured results, then writes the completed output into Notion, Slack, Linear, or Jira.
A company deploys SurfSense on-premises or in a VPC to support controlled research workflows and enterprise identity requirements.
Compare before choosing
Choose self-hosting when infrastructure control and bring-your-own model keys matter more than managed operations.
Choose Pay As You Go when fast hosted access matters and variable per-item or per-page charges fit the workload.
Consider Enterprise when custom connectors, audit logs, SSO, managed updates, or SLA commitments are purchasing requirements.
Features
Cross-source web research
Collect structured material from social platforms, commerce sites, search services, and open web pages through one API or MCP endpoint.
Automated briefs and monitoring
Run scheduled or event-triggered agents that produce briefs, alerts, and ongoing monitoring from current web data.
Private research infrastructure
Self-host the platform, bring model-provider keys, and keep competitive or sensitive research workflows on infrastructure you control.
Pros
- The free open-source edition includes the full platform and lets teams control infrastructure, provider keys, and billing.
- REST API and MCP access let existing applications and agents use the same set of web connectors.
- Hosted calls are billed by returned item or crawled page, and failed calls do not consume credit.
Cons
- Self-hosted users must provide and operate the infrastructure, model keys, and related provider billing.
- Hosted costs vary with the number of returned items and crawled pages, so high-volume workloads require credit monitoring.
- Audit logs, enterprise SSO, SLA commitments, and dedicated support are reserved for a custom Enterprise arrangement.
Pricing Plans
Open Source
- Full platform with connectors, agents, automations, and MCP server
- Unlimited scraping and crawling with user-controlled billing
- Bring your own keys for any model provider
- Community support on Discord
Pay As You Go
- One credit balance for all hosted usage
- REST API and MCP access to platform connectors
- Charges per returned item and crawled page; failed calls are not billed
- Scheduled and event-triggered agents
- Priority support on Discord
Enterprise
- Everything in Pay As You Go
- Custom connectors and agent workflows
- On-premises or VPC deployment
- Audit logs, compliance, and SSO
- Managed updates, SLA commitments, and dedicated support
* Pricing information is for reference only. Please refer to the latest announcement on SurfSense's official website.