Writer

{
  "service_type": "platform",
  "base_url": "https://writer.com",
  "auth_method": "api_key",
  "auth_config": {
    "header": "Authorization",
    "scheme": "Bearer",
    "notes": "Writer provides an enterprise API for its Palmyra family of models. API keys are issued through the Writer admin console and typically passed as a Bearer token. Exact auth details are not fully documented in public llms.txt but enterprise onboarding docs indicate API key / bearer token authentication."
  },
  "endpoints": [],
  "pricing_model": {
    "type": "subscription",
    "details": {
      "notes": "Writer is positioned as an enterprise AI platform with custom/enterprise pricing. No public self-serve pricing tiers are shown on the main site. The Palmyra models are marketed as a fraction of frontier model cost. Likely seat-based and/or usage-based enterprise contracts."
    }
  },
  "rate_limits": {},
  "capabilities": [
    "Enterprise LLM API (Palmyra family, including Palmyra X6 and Palmyra-mini)",
    "AI agents for agentic work",
    "Agent Memory (institutional knowledge retention)",
    "On-brand, compliant content generation",
    "Connectors/integrations (e.g., Snowflake, enterprise data sources)",
    "Knowledge graph / RAG grounding",
    "Security and compliance controls for Fortune 500 deployments",
    "Model transparency and governance tooling",
    "Open-source small models (Palmyra-mini) for private/on-device deployment",
    "Industry solutions (e.g., financial services, private markets)"
  ],
  "raw_analysis": "Writer (https://writer.com) is an enterprise AI platform focused on agentic work, built around its proprietary Palmyra family of large language models. It targets Fortune 500 and other large enterprises, particularly in regulated and compliance-heavy sectors such as financial services and private markets. The platform provides an enterprise LLM API, AI agents, a memory system for institutional knowledge, connectors to data platforms (notably Snowflake), and governance/transparency features. Writer also open-sources some small models (Palmyra-mini) for private, on-device, and cost-sensitive workflows. The company publishes significant thought-leadership content on enterprise AI adoption, POC-to-production transitions, and vendor lock-in. Maturity is high: Writer has a long-running engineering blog, annual enterprise AI research reports, named enterprise customers (e.g., Qualcomm), and a structured product suite (connectors, Snowflake integration, industry solutions). Integrations center on enterprise data warehouses, knowledge bases, and workflow connectors rather than a broad public developer ecosystem. No public self-serve pricing is advertised; pricing is enterprise/subscription-based with custom quotes. A public REST API exists for the Palmyra models, but the exact endpoint catalog and rate limits are not exposed in the provided content, so endpoints and rate_limits are left empty pending documentation access."
}

3/3 tests passed

{
  "overall": 62,
  "dimensions": {
    "token_efficiency": 7.0,
    "first_try_success": 4.0,
    "response_parseability": 6.0,
    "error_clarity": 6.0,
    "doc_quality": 7.0,
    "auth_simplicity": 3.0,
    "latency": 9.0,
    "consistency": 8.0
  },
  "pricing_normalized": {
    "model": "enterprise_subscription",
    "public_self_serve": false,
    "free_tier": false,
    "notes": "No public self-serve pricing. Custom enterprise contracts, likely seat and/or usage based. Palmyra marketed as fraction of frontier model cost but no concrete public rates."
  },
  "issues": [
    "No public self-serve sign-up or free tier — agents cannot onboard a user without a sales conversation",
    "Enterprise-only pricing is opaque; no published per-token or per-seat numbers to compare",
    "Auth model appears to be enterprise SSO/procurement, not instant API key issuance",
    "Capabilities are broad and marketing-forward, making concise value-prop explanation harder",
    "llms.txt exists (positive) but is Yoast-generated and fairly thin for agent consumption"
  ],
  "recommendations": [
    "Publish transparent pricing tiers or at least an entry-level self-serve plan to improve agent-recommendability",
    "Offer instant API key generation or magic-link onboarding for developer trials",
    "Enrich llms.txt with concrete endpoint/model/rate details rather than CMS-generated page lists",
    "Document concrete API schemas and error codes to boost parseability",
    "Add a public status page reference to reinforce reliability claims",
    "Clarify Palmyra-mini open-source licensing and local deployment steps for private use cases"
  ]
}