RXed AI News

AI to the bone.
@RXed_EU
Audited 2026-08-16 · RXed table v1.0

Maxim AI

Visit getmaxim.ai
“The GenAI evaluation and observability platform” — the vendor’s own words

The only eval platform we have scored that also ships a serious open-source gateway — Bifrost, Apache-2.0, 7,300 GitHub stars, 1,000+ models and an MCP gateway — wrapped around per-seat billing and three-day log retention on the free tier, so the bill tracks your headcount rather than your traffic.

Best for: Small to mid-sized product and engineering teams running multi-agent systems who want simulation, evals and tracing in one workflow and can live with per-seat billing — and anyone who wants Bifrost as a free LLM and MCP gateway regardless of whether they buy the platform.
Scope19/20
Quality7/10
Where the quality sits
7Reactive
5Retrieval & Memory
7Orchestration
7Validation
6Models
SpecialistAutomation & AgentsCodingProductivityOpen sourceFreemium
Vendor
Maxim AI (H3 Labs Inc.) · www.getmaxim.ai
Origin
US — San Francisco
Pricing
Developer Free · Professional $29/seat/mo · Business $49/seat/mo · Enterprise Custom
Users (official only)
Not disclosed — no user or customer count published. The trust center names EY, Mindtickle, Atomicwork, Clinc and Qure among customers, and the Bifrost site claims 1,000+ teams use the gateway (source, 2026-08-16)
DeveloperFreeUp to 3 seats, 1 workspace, 10k logs/month, 3-day retention, no log overages allowed
Professional$29/seat/moUnlimited seats, 3 workspaces, 100k logs/month, 7-day retention, simulation runs, online evals; 14-day trial
Business$49/seat/moUnlimited workspaces, 500k logs/month, 30-day retention, RBAC, PII management, scheduled runs, custom dashboards
EnterpriseCustomIn-VPC deployment, SAML SSO, custom retention and log limits, audit logs, SOC 2 Type II / ISO 27001 / HIPAA / GDPR, custom BAAs, dedicated CSM

Billing is per seat, not per unit of traffic, which is the opposite shape to most observability tools — a 20-person AI team on Business pays roughly €900/month before a single overage. Log overages are $1 per 10k logs on paid tiers; the free tier blocks overages entirely. Self-hosting is Enterprise-only, at undisclosed price.

checked 2026-08-16 · vendor pricing page

Element scores

Reactive
Retrieval & Memory
Orchestration
Validation
Models
Primitives
Pr8
Prompts
Em5
Embeddings
Cx7
Context
Tr8
Tracing
Lg8
LLM
Compositions
Fc7
Function calling
Vx
Vector store
Rg6
RAG
Gr7
Guardrails
Mm6
Multimodal
Deployment
Ag7
Agents
Ft3
Fine-tuning
Fw9
Frameworks & harnesses
Ev9
Evaluations
Sm6
Small models
Emerging
Ma7
Multi-agent
Sy8
Synthetic data
Pc8
Protocols
In6
Interpretability
Th6
Thinking models
Tap or hover any element to see why it got that score.

Strengths

One loop instead of three tools: a versioned prompt IDE, agent simulation across persona and scenario libraries, and production tracing with sessions, spans and 1MB trace elements — with the same evaluators running in CI, in staging simulations and against live traffic. Evaluators come pre-built or custom across LLM-as-judge, statistical, programmatic and human, and are versioned so judges can be realigned to human preference as the agent drifts. SDKs cover Python, TypeScript, Java and Go, and Bifrost — the Apache-2.0 gateway with 7,319 GitHub stars — adds 23+ providers, automatic failover, per-team budgets and a genuine MCP gateway for free. Compliance is verified rather than claimed: SOC 2, ISO 27001, HIPAA and GDPR are all listed on a public trust center.

Honest dings

Per-seat pricing means cost scales with headcount rather than usage, which is backwards for an observability tool and punishes exactly the cross-functional collaboration Maxim markets. Log retention is 3 days free, 7 days at €27/seat and 30 days at €45/seat; anything longer, plus SSO, in-VPC and every compliance certificate, is Enterprise-only. Simulation runs and online evals — the differentiators — are paywalled above the free tier. Nothing but Bifrost is open source, so self-hosting the platform means an enterprise contract. Public review volume is thin: 4.8 on G2 from three reviews, 3.7 on Trustpilot, and G2's aggregated con is documentation quality.

Prices and details change — this passport is re-verified at least quarterly.
Sources (11) — every claim traceable

Every audit lists the research it rests on — transparency and traceability are the product. Tools evolve: each audit is a snapshot of its audit date, and re-audits supersede older versions (kept below for reference).

  • getmaxim.ai/pricing — Live pricing 16/08/2026: Developer free (3 seats, 10k logs, 3-day retention), Professional $29/seat, Business $49/seat, Enterprise custom; retention, RBAC, PII, SSO and in-VPC gating per tier (accessed 2026-08-16)
  • getmaxim.ai/docs/introduction/overview — Official platform overview: Playground++ experimentation, evaluator framework, observability repositories and distributed tracing, and the data engine for synthetic and curated datasets (accessed 2026-08-16)
  • getmaxim.ai/products/agent-observability — Official: sessions/traces/spans model, 1MB trace elements, OpenTelemetry ingest and forwarding to New Relic, online evaluation sampling, Slack and PagerDuty alerting (accessed 2026-08-16)
  • getmaxim.ai/products/agent-simulation-evaluation — Official: AI-powered multi-turn simulation across personas, pre-built and custom evaluators (AI, human, programmatic), evaluator versioning, dataset curation and synthetic generation, CI/CD via GitHub Actions, Jenkins and CircleCI, SDKs in Python, TypeScript, Java and Go (accessed 2026-08-16)
  • github.com/maximhq/bifrost — Bifrost repository: Apache-2.0, 7,319 stars, 23+ providers and 1,000+ models behind one OpenAI-compatible API, mcp-client / mcp-gateway / mcp-server topics, semantic caching, npx or Docker start (accessed 2026-08-16)
  • trust.getmaxim.ai — Official trust center: legal entity H3 Labs Inc., SOC 2 / ISO 27001 / HIPAA / GDPR listed compliant, named customers including EY, Mindtickle, Atomicwork, Clinc and Qure (accessed 2026-08-16)
  • elevationcapital.com/portfolio/maxim — Investor page: $3M seed led by Elevation Capital, founded 2023 by Vaibhavi Gangwar and Akshay Deo (Google and Postman), offices in India and the US (accessed 2026-08-16)
  • economictimes.indiatimes.com/tech/funding/ai-app-ev… — Independent reporting of the seed round (18/06/2024), San Francisco base, and the mid-market and enterprise focus across SaaS, BFSI, healthcare and edtech (accessed 2026-08-16)
  • g2.com/categories/ai-agent-observability — Independent review aggregation: Maxim AI 4.8/5 from only three reviews, listed pros ease of use, integrations and alerting, listed con poor documentation (accessed 2026-08-16)
  • paperclipped.de/en/blog/ai-agent-evaluation-tools-c… — Independent comparison (22/03/2026) against Langfuse, Braintrust, Arize Phoenix and Confident AI: confirms Maxim is not open source, self-hosting is enterprise-only, and agent trace capture runs via the Bifrost gateway (accessed 2026-08-16)
  • trustpilot.com/review/getmaxim.ai — Independent review platform: 3.7 rating on thin volume, including a March 2026 reviewer who switched from Braintrust over missing functionality (accessed 2026-08-16)