CoreBase

Governed AI-agent infrastructure that safely connects LLMs to your real data systems

Freemium· Free (1 source, $1 credit) / Pro $49 per mo (5 sources, $5 AI credit) / Team $149 per mo (unlimited sources, $15 AI credit) / Enterprise custom (air-gapped)AgentsBring-your-own-key: Claude, GPT-4/4o, Gemini (plus built-in fallback models)

Visit website →

Best for

Platform and data teams at regulated or mid-market companies that need to deploy LLM agents against production databases and SaaS tools with real audit, RBAC and approval controls.

Skip if

Solo hackers, prototype builders or consumer chatbot projects where governance, RBAC and audit logging are pure overhead rather than requirements.

CoreBase is an agent-infrastructure platform aimed squarely at the awkward middle of enterprise AI: teams that want LLM-powered assistants working against their production databases, SaaS tools, and legacy systems, but cannot afford the free-for-all that most prompt-your-warehouse products invite. It ships as a governance layer on top of 50+ pre-built connectors (PostgreSQL, MSSQL, Salesforce, Slack, Jira and similar), a zero-trust outbound-only gateway called CoreMCP for reaching on-premises systems without opening inbound ports, and a policy engine that enforces read-only defaults, per-user isolation, approval gates for write actions, RBAC, and a complete audit trail of every prompt, tool call and row returned.

The headline feature is Query Memory, which lets the agent progressively learn schema, joins and business terminology across sessions so analysts stop re-explaining what "active customer" means. Attack-screening intercepts prompt injection attempts before they reach connected systems, and scoped credentials can be revoked without redeploying anything. Developers get a REST API and an embeddable chat widget on every paid tier, and enterprises can run the whole stack air-gapped.

Model choice is bring-your-own-key for Claude, GPT and Gemini, with built-in fallback credits included in each plan. Typical workflows are natural-language data questions against governed warehouses, internal-tool copilots that can safely take actions in Salesforce or Jira, on-call agents that read logs and open tickets, and compliance-friendly chat interfaces that need to prove exactly what the LLM saw and did. It is less a chatbot and more the boring plumbing that makes chatbots deployable inside regulated companies.

Editor's take

CoreBase is what you reach for when the security team has already vetoed your first two agent prototypes. The zero-trust outbound bridge, read-only defaults and open-source CoreMCP are the right primitives, and BYO-key keeps you off a single model. The catch is the middle tiers feel tight — most serious teams will end up on Team or Enterprise fast — but for regulated shops that's still cheaper than rolling your own guardrails.

— The AI Tool Bible editorial team

Pros

✅ Zero-trust CoreMCP bridge means on-prem systems stay behind the firewall — no inbound ports opened
✅ Read-only by default with approval gates on writes, so a jailbroken agent cannot drop a table
✅ Full audit trail of prompts, tool calls, rows returned — genuinely useful for SOC2/HIPAA reviews
✅ 50+ prebuilt connectors including MSSQL and legacy SaaS most agent frameworks ignore
✅ Bring-your-own-key for Claude, GPT and Gemini — no vendor model lock-in
✅ CoreMCP bridge is open source on GitHub, so the sensitive side is inspectable
✅ Query Memory reduces the schema-re-explaining that plagues text-to-SQL agents
✅ Air-gapped deployment path exists for regulated buyers

Cons

⚠️ Governance overhead is overkill for solo builders or hobby projects
⚠️ AI credits on lower tiers ($1–$15/mo) run out fast on real workloads once BYO-key isn't set
⚠️ Only 5 data sources on the $49 Pro tier — mid-size teams jump straight to Team or Enterprise
⚠️ Ecosystem is young; fewer community integrations than LangChain/LlamaIndex-based stacks
⚠️ Air-gapped and advanced policy controls are gated behind Enterprise custom pricing

Use cases

Governed text-to-SQL over production warehousesInternal Slack copilot with Salesforce and Jira actionsOn-call agent that reads logs and opens ticketsCustomer-support assistant grounded in on-prem CRM dataCompliance-audited AI chat for regulated industriesEmbedded AI assistant inside SaaS products via widgetRead-only analyst copilot with approval gates on writesAir-gapped LLM deployment for defense or healthcare