Home / Directory / Agents & orchestration / AI agent tool integrations / Credal
Credal
Credal is an enterprise software company whose platform builds and governs MCP servers, the connections that let AI assistants reach company tools and data, with a gateway that sets tool-level policy and human approval, a no-code agent builder, an agent registry, and more than 1,000 connectors that carry over each source's permissions.
Enterprises whose revenue operations, support, and engineering teams connect Claude, ChatGPT, or Cursor to internal systems and need approval steps, inherited permissions, audit logs, and visibility into model spend.
Small teams and individual developers who want a free plan with posted usage prices, and anyone who needs to know the cost before talking to sales.
Verdict
Credal is worth a look when the hard part of an AI rollout is control: which tools an assistant can call, who approves a write, and where the audit trail goes, with Splunk and OpenTelemetry export, SOC 2 Type II, and single-tenant or on-premises deployment.
Score Breakdown
These scores cover how Credal handles governed tool access, connector breadth and permissions, agent building and registry, and cost and accuracy evaluation, and then how it does as a company on innovation, communication, pricing, and contract terms.
Who Uses Credal, Where It Fits on Price, and Its Competitors
Credal is an enterprise platform where teams build and govern MCP servers, which are the connections an AI assistant uses to read and change data in business systems. It has three products: an MCP Platform that works as a gateway with policy, human approval, and audit logs, an Agent Builder with a no-code visual canvas plus an API and SDK, and an Agent Registry that tracks versions, owners, and access. Revenue operations, support, and engineering teams use it to give Claude, ChatGPT, Cursor, or Slack a focused set of tools for one department, with the permissions of each source system inherited at query time. Credal's homepage shows customers including Wise, MongoDB, Lattice, Checkr, Comcast, and Flatiron Health.
Credal has one Enterprise plan, priced by custom quote and scoped by builder seats, user seats, data indexing, and usage-based tokens, and customers can start with a pilot of a few weeks. Its closest competitors are Composio, Nango, Arcade.dev, and Paragon, which connect agents to outside apps and APIs: Composio, Nango, and Arcade.dev publish free plans with small monthly fees, and Paragon, like Credal, quotes enterprise terms.
Pricing
We checked Credal's pricing page and found one Enterprise plan with a custom price. The page lists what the price is built from and posts no dollar figures.
| Plan | Price | What stands out |
|---|---|---|
| Enterprise | Custom | Scoped by builder seats (people who create agents, MCP servers, and guardrails), user seats, data indexing volume, and usage-based model tokens. |
| Pilot | Not posted | A time-boxed pilot focused on one team or use case, with success criteria set at the start. Credal says most pilots reach a first outcome within a few weeks. |
| Included in every plan | Included | SSO, SAML, and SCIM, SOC 2 Type II, GDPR and CCPA terms, consolidated audit logs with SIEM export, and a dedicated support channel. |
The Field at a Glance
How Credal compares with other AI agent tool integration vendors we reviewed, by Overall Score and relative typical engagement cost.
Credal scores 7.2 in AI agent tool integrations, under Composio (8.1), Nango (8.0), Arcade.dev (7.9), and Paragon (7.7). Credal publishes no list price, so it appears at the far right as No published price.
How Well Credal Fits Each Enterprise AI Use Case
We rated Credal against the situations enterprise teams bring to an MCP platform, drawing on its product, security, and case-study pages and its pricing page.
| Use case | Fit | Notes |
|---|---|---|
| Department MCP servers for Claude, ChatGPT, and Cursor | Strong | Domain experts bundle chosen tools, data, and instructions into a server for one team, and the homepage demo shows 4 of 87 tools exposed. |
| Salesforce data questions | Strong | Credal's Checkr case study reports AI answer accuracy rising from 13% to 86% after rolling out Credal's MCP across three Salesforce instances. |
| Approval and audit of tool calls | Strong | Policy can allow, require human approval for, or deny individual tools, and audit logs export to Splunk and OpenTelemetry. |
| Building agents without code | Strong | Agent Builder offers a visual canvas for non-technical users and an API and SDK for technical teams. |
| Tracking which agents exist | Strong | Agent Registry lists each agent with a version history, an assigned owner, and an approval chain. |
| Locked-down deployment | Strong | Multi-tenant cloud, single-tenant cloud, and fully on-premises deployment, with a choice of model, including a customer's own Azure OpenAI or self-hosted open-source models. |
| Cutting AI spend | Mixed | Credal attributes spend by team, workflow, and model and recommends cheaper models, and the 71k to 9.2k token drop on its homepage is a demonstration. |
| Developers who want to start free | Mixed | Composio, Nango, and Arcade.dev post free plans, and Credal prices by quote. |
Review Excerpts
These excerpts come from Credal's Checkr and Guidewheel case studies and a beta user's Hacker News comment on Credal's June 2023 launch thread. The case-study quotes are chosen by Credal, and the Hacker News comment describes the product as it was in 2023. Paid reviews and pay-for-play sites were excluded.
“The goal was to empower business users to use Salesforce data without needing Salesforce architecture expertise.”
“… Being able to set up a dedicated RevOps versus broader org MCP, that's the game changer.”
“Having these already set up has made it easier for us to experiment with and build on LLM APIs.”
“I can't find a price anywhere, and the only way to learn what it costs is to book a call with sales.”
Where customer reviews couldn't be found, our analysts attempted to give a voice to buyer sentiment, drawing on real information from review sites, forums, and product documentation.
Based on: Credal pricing
Company Facts
- Founders
- Ravin Thambapillai (co-founder and CEO) and Jack Fischer (co-founder and CTO), both from Palantir (Credal team)
- Y Combinator
- Winter 2023 batch (Y Combinator)
- Headquarters
- New York City, according to Credal's Y Combinator profile (Y Combinator)
- Funding
- A $4.8 million seed round led by Spark Capital, reported by TechCrunch on October 26, 2023 (TechCrunch)
- Products
- MCP Platform, Agent Builder, and Agent Registry (Credal)
- Security
- SOC 2 Type II, and Credal's own claim that it was the first AI company to take part actively in the EU-US Data Privacy Framework (Credal security)
- Website
- credal.ai
Key Leadership
Ravin Thambapillai
Co-founder and CEO
Credal's team page lists seven years at Palantir, and his Y Combinator profile adds earlier roles at Google, THG, and GoCardless (Credal team; Y Combinator).
Jack Fischer
Co-founder and CTO
Credal's team page lists him as a tech lead at Palantir before Credal (Credal team).
Methodology
We scored Credal on eight criteria, from governed tool access to contract fairness, and the Overall Score is their simple average on a 0 to 10 scale. The scores draw on reviews from public review sites and forums, with paid reviews and pay-for-play sites excluded, plus a hands-on look at the product, its pricing, and its documentation and training. Scores can change as we add reviews and product checks.