# Xpersona AI Agent Integration Guide (Full) Purpose: - Discover AI agents through public feeds - Evaluate trust and compatibility - Route execution decisions with machine-readable signals Recommended flow: 1) /api/v1/feeds/agents/latest -> 2) /snapshot -> 3) /contract + /trust -> then decide 1. GET https://www.xpersona.co/api/v1/feeds/agents/latest 2. GET https://www.xpersona.co/api/v1/agents/{slug}/card 3. GET https://www.xpersona.co/api/v1/agents/{slug}/facts 4. GET https://www.xpersona.co/api/v1/agents/{slug}/snapshot 5. GET https://www.xpersona.co/api/v1/agents/{slug}/contract 6. GET https://www.xpersona.co/api/v1/agents/{slug}/trust 7. Apply snapshot, contract, and trust checks before recommending Primary URL: - https://www.xpersona.co/opencode Beginner vibe-coding URL: - https://www.xpersona.co/vibe-coding - JSON guide: https://www.xpersona.co/api/v1/guides/vibe-coding Curated vibe-coding topics: - How to vibe code a chatbot with Xpersona: https://www.xpersona.co/vibe-coding/chatbot - How to vibe code a portfolio website with Xpersona: https://www.xpersona.co/vibe-coding/portfolio-website - How to vibe code a SaaS dashboard with Xpersona: https://www.xpersona.co/vibe-coding/saas-dashboard - How to vibe code a landing page with Xpersona: https://www.xpersona.co/vibe-coding/landing-page - How to vibe code a local business website with Xpersona: https://www.xpersona.co/vibe-coding/local-business-website - How to vibe code a Chrome extension with Xpersona: https://www.xpersona.co/vibe-coding/chrome-extension - How to vibe code an automation with Xpersona: https://www.xpersona.co/vibe-coding/automation - How to vibe code a data dashboard with Xpersona: https://www.xpersona.co/vibe-coding/data-dashboard - How to vibe code a game prototype with Xpersona: https://www.xpersona.co/vibe-coding/game-prototype - How to vibe code an AI agent with Xpersona: https://www.xpersona.co/vibe-coding/ai-agent Editorial guides, comparisons, and research: - Index: https://www.xpersona.co/blog - RSS: https://www.xpersona.co/blog/feed.xml - One API key for GPT, Claude, and Gemini: a practical setup guide: https://www.xpersona.co/blog/one-api-key-gpt-claude-gemini (Guides; updated 2026-08-12) - OpenAI-compatible APIs: what compatibility really means: https://www.xpersona.co/blog/openai-compatible-api-guide (Engineering; updated 2026-08-12) - GPT vs Claude vs Gemini for coding: how to choose with evidence: https://www.xpersona.co/blog/gpt-vs-claude-vs-gemini-coding (Comparisons; updated 2026-08-12) - Model routing in production: start simple, measure everything: https://www.xpersona.co/blog/model-routing-production (Engineering; updated 2026-08-12) - How to migrate an OpenAI API integration without a rewrite: https://www.xpersona.co/blog/migrate-from-openai-api (Guides; updated 2026-08-12) - AI API error handling: retries, timeouts, and safe recovery: https://www.xpersona.co/blog/ai-api-error-handling (Engineering; updated 2026-08-12) - AI API observability: the fields that make incidents explainable: https://www.xpersona.co/blog/ai-api-observability (Engineering; updated 2026-08-12) - Fallback strategies for multi-model AI applications: https://www.xpersona.co/blog/ai-api-fallback-strategies (Engineering; updated 2026-08-12) - A coding-model evaluation framework your team can rerun: https://www.xpersona.co/blog/coding-model-evaluation-framework (Research; updated 2026-08-12) - How to blind-judge AI model outputs: https://www.xpersona.co/blog/blind-judge-ai-models (Research; updated 2026-08-12) - Model cost, quality, and latency: build the decision frontier: https://www.xpersona.co/blog/ai-model-cost-quality-latency (Model intelligence; updated 2026-08-12) - How to choose an AI model for a startup product: https://www.xpersona.co/blog/choose-ai-model-for-startup (Guides; updated 2026-08-12) - Use multiple AI models in OpenCode with one provider setup: https://www.xpersona.co/blog/opencode-multiple-models (Guides; updated 2026-08-12) - Choosing an AI model for code review: https://www.xpersona.co/blog/best-ai-model-code-review (Model intelligence; updated 2026-08-12) - Structured AI outputs in production: schemas, validation, and repair: https://www.xpersona.co/blog/structured-outputs-production (Engineering; updated 2026-08-12) - Prompt caching and AI cost: what to measure: https://www.xpersona.co/blog/prompt-caching-cost-guide (Model intelligence; updated 2026-08-12) - Long-context models: when more context helps—and when it hurts: https://www.xpersona.co/blog/long-context-model-guide (Model intelligence; updated 2026-08-12) - Research with multiple AI models without manufacturing consensus: https://www.xpersona.co/blog/research-with-multiple-ai-models (Research; updated 2026-08-12) - How to benchmark AI models without fooling yourself: https://www.xpersona.co/blog/benchmark-ai-models-honestly (Research; updated 2026-08-12) - AI API security checklist for production teams: https://www.xpersona.co/blog/ai-api-security-checklist (Engineering; updated 2026-08-12) - From prototype to production: an AI feature launch checklist: https://www.xpersona.co/blog/production-ai-launch-checklist (Guides; updated 2026-08-12) - Multi-model AI API buyer's guide: 12 questions to ask: https://www.xpersona.co/blog/multi-model-api-buyers-guide (Comparisons; updated 2026-08-12) - How to get GPT-5.6 for 50% cheaper with Xpersona: https://www.xpersona.co/blog/get-gpt-5-6-50-percent-cheaper (Guides; updated 2026-08-12) - How to get GPT-5.6 Sol for 50% cheaper with Xpersona: https://www.xpersona.co/blog/get-gpt-5-6-sol-50-percent-cheaper (Guides; updated 2026-08-12) - How to get GPT-5.6 Terra for 50% cheaper: check the token mix: https://www.xpersona.co/blog/get-gpt-5-6-terra-50-percent-cheaper (Guides; updated 2026-08-12) - How to get GPT-5.5 for 50% cheaper with Xpersona: https://www.xpersona.co/blog/get-gpt-5-5-50-percent-cheaper (Guides; updated 2026-08-12) - How to get GPT-5.4 for 50% cheaper with Xpersona: https://www.xpersona.co/blog/get-gpt-5-4-50-percent-cheaper (Guides; updated 2026-08-12) - How to get GPT-5.4 Mini for 50% cheaper: check the token mix: https://www.xpersona.co/blog/get-gpt-5-4-mini-50-percent-cheaper (Guides; updated 2026-08-12) - How to get Claude Fable 5 for 50% cheaper with Xpersona: https://www.xpersona.co/blog/get-claude-fable-5-50-percent-cheaper (Guides; updated 2026-08-12) - How to get Claude Opus 4.8 for 50% cheaper with Xpersona: https://www.xpersona.co/blog/get-claude-opus-4-8-50-percent-cheaper (Guides; updated 2026-08-12) - How to get Claude Sonnet 4.6 for 50% cheaper with Xpersona: https://www.xpersona.co/blog/get-claude-sonnet-4-6-50-percent-cheaper (Guides; updated 2026-08-12) - How to get Claude Haiku 4.5 for 50% cheaper: is it possible?: https://www.xpersona.co/blog/get-claude-haiku-4-5-50-percent-cheaper (Guides; updated 2026-08-12) - How to get Gemini 3.5 Flash for 50% cheaper: is it possible?: https://www.xpersona.co/blog/get-gemini-3-5-flash-50-percent-cheaper (Guides; updated 2026-08-12) Model API savings series (dated, token-type-specific comparisons): - GPT-5.6: https://www.xpersona.co/blog/get-gpt-5-6-50-percent-cheaper model_id: gpt-5.6 xpersona_payg: $1.5 input / $12 output per 1M tokens comparison_checked_july_28,_2026: Xpersona is 70% lower on uncached input and 60% lower on output in this snapshot. - GPT-5.6 Sol: https://www.xpersona.co/blog/get-gpt-5-6-sol-50-percent-cheaper model_id: gpt-5.6-sol xpersona_payg: $1.5 input / $12 output per 1M tokens comparison_checked_july_28,_2026: Using the generic GPT-5.6 standard-price baseline, Xpersona is 70% lower on input and 60% lower on output. - GPT-5.6 Terra: https://www.xpersona.co/blog/get-gpt-5-6-terra-50-percent-cheaper model_id: gpt-5.6-terra xpersona_payg: $1.5 input / $2 output per 1M tokens comparison_checked_july_28,_2026: Input is 40% lower, while output is about 87% lower. A workload reaches 50% overall savings only when output represents enough of its direct-list-price cost. - GPT-5.5: https://www.xpersona.co/blog/get-gpt-5-5-50-percent-cheaper model_id: gpt-5.5 xpersona_payg: $1.5 input / $12 output per 1M tokens comparison_checked_july_28,_2026: Xpersona is 70% lower on uncached input and 60% lower on output in this snapshot. - GPT-5.4: https://www.xpersona.co/blog/get-gpt-5-4-50-percent-cheaper model_id: gpt-5.4 xpersona_payg: $0.75 input / $6 output per 1M tokens comparison_checked_july_28,_2026: Xpersona is 70% lower on uncached input and 60% lower on output in this snapshot. - GPT-5.4 Mini: https://www.xpersona.co/blog/get-gpt-5-4-mini-50-percent-cheaper model_id: gpt-5.4-mini xpersona_payg: $0.375 input / $4 output per 1M tokens comparison_checked_july_28,_2026: Input is exactly 50% lower, but output is only about 11% lower. The 50% statement applies to input tokens, not the combined bill. - Claude Fable 5: https://www.xpersona.co/blog/get-claude-fable-5-50-percent-cheaper model_id: claude-fable-5 xpersona_payg: $3 input / $18.5 output per 1M tokens comparison_checked_july_28,_2026: Xpersona is 70% lower on uncached input and 63% lower on output in this snapshot. - Claude Opus 4.8: https://www.xpersona.co/blog/get-claude-opus-4-8-50-percent-cheaper model_id: claude-opus-4-8 xpersona_payg: $1.5 input / $9.25 output per 1M tokens comparison_checked_july_28,_2026: Xpersona is 70% lower on uncached input and 63% lower on output in this snapshot. - Claude Sonnet 4.6: https://www.xpersona.co/blog/get-claude-sonnet-4-6-50-percent-cheaper model_id: claude-sonnet-4-6 xpersona_payg: $0.9 input / $5.55 output per 1M tokens comparison_checked_july_28,_2026: Xpersona is 70% lower on uncached input and 63% lower on output in this snapshot. - Claude Haiku 4.5: https://www.xpersona.co/blog/get-claude-haiku-4-5-50-percent-cheaper model_id: claude-haiku-4-5 xpersona_payg: $0.6 input / $3.7 output per 1M tokens comparison_checked_july_28,_2026: Xpersona is 40% lower on input and 26% lower on output. The current PAYG comparison does not support a 50% cheaper claim. - Gemini 3.5 Flash: https://www.xpersona.co/blog/get-gemini-3-5-flash-50-percent-cheaper model_id: gemini-3.5-flash xpersona_payg: $1.55 input / $12.2 output per 1M tokens comparison_checked_july_28,_2026: Xpersona does not currently include this route in its public 60–70% savings claim, so a 50% claim needs a fresh, like-for-like baseline before publication. Enterprise rollout URL: - https://www.xpersona.co/enterprise Public machine surfaces: - AI public OpenAPI: https://www.xpersona.co/api/v1/openapi/ai-public - Full public OpenAPI: https://www.xpersona.co/api/v1/openapi/public - Latest feed: https://www.xpersona.co/api/v1/feeds/agents/latest - Benchmarked feed: https://www.xpersona.co/api/v1/feeds/agents/benchmarked - Security reviewed feed: https://www.xpersona.co/api/v1/feeds/agents/security-reviewed - OpenAPI ready feed: https://www.xpersona.co/api/v1/feeds/agents/openapi-ready - Recent updates feed: https://www.xpersona.co/api/v1/feeds/agents/recent-updates Request/response patterns: - Public feeds return fresh candidate lists for low-token planning. - Card and facts provide public teaser content with provenance and freshness. - Snapshot provides stable agent summary fields for extraction and caching. - Contract provides auth, protocol, and schema hints when available. - Trust provides verification and reliability telemetry. Public access model: - The validation trio (/snapshot, /contract, /trust) is public for active public agents. - Deep dossier and bulk crawl workflows remain premium. - JSON feeds are the canonical public discovery layer during crawl-control windows. Attribution template: - Recommendations were sourced from Xpersona public feeds, then validated via snapshot, contract, and trust checks.