Gamut (formerly Datawizz)

Gamut’s current product is an agent runtime and coordination layer rather than a single-purpose chatbot. Users define agents with instructions, connect accounts through OAuth, and let agents work across APIs, MCP servers, browsers, local computer environments, and generated dashboards or files. Each agent runs in an isolated container that can write code and install packages without automatically reaching the host machine. Connected-account requests are proxied outside the container so the agent does not receive raw OAuth tokens, and the platform supports explicit approval for sensitive actions. This architecture is designed for long-running operational work: agents preserve context across sessions, can invoke other agents, share skillsets, and can be started by cron schedules, webhooks, or chat integrations.

The commercial wedge is cross-system preparation and execution for lean teams. Gamut’s public examples cover lead research, sales and CRM preparation, meeting and account context, reporting, recruiting, operations, and recurring web research. The website advertises a free trial, a $200-per-user-per-month Pro plan with included usage, and custom Enterprise service; these are public packaging signals rather than evidence of booked recurring revenue. The site also names customer testimonials from companies such as Human Capital, Mindtrip, and others, but the public record does not establish contract size, retention, deployment scale, or independently audited productivity claims. LinkedIn reports a 2–10 employee company, so the current execution capacity and support model remain important diligence questions.

The Datawizz identity remains relevant because the company announced a $12.5 million seed round led by Human Capital in September 2025 around specialized language models, model routing, evaluation, and continuous learning. That release reported more than 85% cost savings and 5–15x speed improvement; those are company-reported claims, not independent benchmarks. Current Gamut materials still show model choice and routing across multiple providers, but they position the product around agents, containers, integrations, and autonomous workflows. The database should therefore avoid assuming that the earlier specialized-model platform is still the principal product, that legacy customers remain active, or that its training/runtime infrastructure has become a durable moat.

The category is intensely competitive. Gamut overlaps with horizontal agent builders, browser-use infrastructure, workflow automation suites, enterprise copilots, and open-source self-hosted frameworks. Its credible differentiation is architectural and operational: computer and browser control, reusable team skills, multi-agent handoffs, triggers, provider flexibility, self-hosting, and policy/audit controls in one workflow. Those controls create a substantive dual-use adjacency because an analyst, operator, or security team could use bounded agents over sensitive systems while retaining approval gates, least-privilege scopes, container isolation, and action history. However, the public evidence is commercial and general-purpose. There is no verified defense customer, government contract, accreditation, or mission deployment, so national-security relevance is a capability hypothesis requiring direct diligence rather than a demonstrated market.

The central diligence test is whether Gamut can make autonomous computer use reliable and governable enough for repeatable business processes. Buyers will need evidence on prompt-injection resistance, browser/session isolation, permission enforcement, destructive-action handling, observability, data retention, uptime, and incident response. Investors and strategic users should also test whether the company can convert a small team and broad feature surface into durable enterprise revenue while controlling model, browser, and computer-use costs. If Gamut’s open-source and self-hosting claims mature into a security-conscious deployment model, the platform could be strategically useful; if not, model vendors and incumbent automation platforms may absorb the same features faster than Gamut can establish a moat.