{"id":2,"date":"2024-07-30T10:40:00","date_gmt":"2024-07-30T10:40:00","guid":{"rendered":"https:\/\/celso.world\/noograms\/?page_id=2"},"modified":"2026-02-06T03:51:11","modified_gmt":"2026-02-06T03:51:11","slug":"ai-agents","status":"publish","type":"page","link":"https:\/\/celso.world\/noograms\/ai-agents\/","title":{"rendered":"AI Agents"},"content":{"rendered":"\n
\"\"
NooGrams, AI Gen based on illustration + prompt by Celso singo Aramaki (July 2024)<\/figcaption><\/figure>\n\n\n\n

The Use of AI Agents in Noograms<\/h1>\n\n\n\n

By Celso Singo Aramaki + AI (updated: January 2026)<\/em><\/p>\n\n\n\n

Introduction<\/h3>\n\n\n\n

Between 2024 and early 2026, \u201cAI agents\u201d shifted from a buzzword to a more practical set of tools: LLMs connected to retrieval, structured memory, and external tools<\/strong>, coordinated through workflows and evaluation checks. In creative production, the most reliable use is not autonomous storytelling, but assisted development<\/strong>\u2014helping authors test scenes, maintain continuity, and iterate faster without losing control of voice and intent.<\/p>\n\n\n\n

In Noograms, AI agents are used as a development layer around the graphic novel: a way to support character consistency, research, dialogue exploration, and narrative structure.<\/p>\n\n\n\n

\n\n\n\n

What \u201cAI agents\u201d means in 2026<\/h3>\n\n\n\n

In this context, an agent is best understood as:<\/p>\n\n\n\n

    \n
  • a language model<\/strong> guided by role instructions (e.g., \u201ccharacter voice tester,\u201d \u201ccontinuity checker,\u201d \u201ceditor\u201d)<\/li>\n\n\n\n
  • connected to retrieval<\/strong> (so it references the project\u2019s canon instead of guessing)<\/li>\n\n\n\n
  • equipped with tool use<\/strong> (structured functions for summarizing, cross-checking, outlining, tagging, and compiling)<\/li>\n\n\n\n
  • constrained by guardrails<\/strong> (what it is allowed to claim, and how it should respond when uncertain)<\/li>\n<\/ul>\n\n\n\n

    This is less about \u201ccharacters becoming alive\u201d and more about making a long-form project easier to manage<\/strong>.<\/p>\n\n\n\n

    \n\n\n\n

    AI agents as character workbenches (not replacements)<\/h3>\n\n\n\n

    Each major character can be represented as an \u201cagent profile\u201d built from structured inputs:<\/p>\n\n\n\n

      \n
    • backstory and timeline<\/li>\n\n\n\n
    • personality constraints and voice notes<\/li>\n\n\n\n
    • relationships and conflicts<\/li>\n\n\n\n
    • beliefs, goals, and contradictions<\/li>\n\n\n\n
    • known facts (\u201ccanon\u201d) vs. unknowns (\u201copen questions\u201d)<\/li>\n<\/ul>\n\n\n\n

      Agents then help stress-test the writing process by generating controlled variants<\/strong> of dialogue and reactions under specific scene constraints. Outputs are treated as drafts and prompts for revision\u2014not final character truth.<\/p>\n\n\n\n

      \n\n\n\n

      Learning and evolution: what is real vs. what is simulated<\/h3>\n\n\n\n

      In 2026, it\u2019s possible to create the feeling<\/em> of character evolution through:<\/p>\n\n\n\n

        \n
      • updating the character\u2019s state<\/strong> after key events (what changed, what they learned, what they now avoid)<\/li>\n\n\n\n
      • adding new canon notes to the retrieval store<\/li>\n\n\n\n
      • tracking relationship dynamics (trust, tension, dependency) as structured variables<\/li>\n<\/ul>\n\n\n\n

        However, this evolution is typically curated<\/strong>, not automatic. Without careful constraints, \u201cself-learning\u201d behaviors drift, contradict earlier canon, or inflate drama in ways that don\u2019t fit the story. For Noograms, evolution is handled through author-approved updates<\/strong> to character state and canon.<\/p>\n\n\n\n

        \n\n\n\n

        Inter-agent relationships and scene simulation<\/h3>\n\n\n\n

        Multi-agent setups are useful for one thing: testing interaction patterns<\/strong>.<\/p>\n\n\n\n

        A controlled \u201cscene simulation\u201d can run with:<\/p>\n\n\n\n

          \n
        • a fixed setting and stakes<\/li>\n\n\n\n
        • strict scene objectives (what must happen)<\/li>\n\n\n\n
        • turn limits<\/li>\n\n\n\n
        • grounded relationship states<\/li>\n\n\n\n
        • a moderator agent that flags contradictions, tone breaks, or implausible shifts<\/li>\n<\/ul>\n\n\n\n

          The output becomes a workbench transcript<\/strong>\u2014useful for discovering tension lines, dialogue rhythms, and alternative beats\u2014without deciding the final page for the author.<\/p>\n\n\n\n

          \n\n\n\n

          Technical implementation (practical stack)<\/h3>\n\n\n\n

          A typical 2026 implementation for this kind of work involves:<\/p>\n\n\n\n

            \n
          • RAG (Retrieval-Augmented Generation):<\/strong> a searchable canon store (characters, timeline, locations, themes, prior chapters)<\/li>\n\n\n\n
          • Structured memory:<\/strong> scene-level state updates and relationship variables stored explicitly<\/li>\n\n\n\n
          • Tool calling \/ function interfaces:<\/strong> repeatable actions like \u201csummarize character arc,\u201d \u201ccheck continuity,\u201d \u201cgenerate 5 dialogue options in voice,\u201d \u201ccompare scene variants,\u201d \u201cextract beats\u201d<\/li>\n\n\n\n
          • Workflow orchestration:<\/strong> agent pipelines (writer \u2192 editor \u2192 continuity \u2192 style) with logging<\/li>\n\n\n\n
          • Evaluations and checklists:<\/strong> lightweight tests for contradictions, tone drift, verbosity, or character voice mismatch<\/li>\n<\/ul>\n\n\n\n

            This architecture favors repeatability<\/strong> over novelty. The point is consistency and speed, not spectacle.<\/p>\n\n\n\n

            \n\n\n\n

            Role of agents in narrative development<\/h3>\n\n\n\n

            Agents can support narrative development in three grounded ways:<\/p>\n\n\n\n

              \n
            1. Continuity discipline<\/strong>
              Catch timeline errors, inconsistent motivations, and relationship contradictions early.<\/li>\n\n\n\n
            2. Iteration speed<\/strong>
              Generate structured alternatives (dialogue variants, beat permutations, scene openings) quickly, then select and rewrite.<\/li>\n\n\n\n
            3. Decision support<\/strong>
              Provide comparisons: what changes if a scene is moved, shortened, or reframed; what consequences ripple into later chapters.<\/li>\n<\/ol>\n\n\n\n

              The plot remains authored. Agents help explore options and reduce friction in revision.<\/p>\n\n\n\n

              \n\n\n\n

              Work<\/h3>\n\n\n\n

              In Noograms, AI agents are used as a practical creative layer: canon-aware assistants that help test scenes, maintain character integrity, and accelerate iteration<\/strong>. The approach is deliberately conservative: constrained inputs, retrieval grounding, explicit state updates, and human editorial control. If the system adds value, it does so quietly\u2014by making long-form storytelling more maintainable and more coherent over time.<\/p>\n\n\n\n

              \n","protected":false},"excerpt":{"rendered":"

              The Use of AI Agents in Noograms By Celso Singo Aramaki + AI (updated: January 2026) Introduction Between 2024 and early 2026, \u201cAI agents\u201d shifted from a buzzword to a more practical set of tools: LLMs connected to retrieval, structured memory, and external tools, coordinated through workflows and evaluation checks. In creative production, the most […]<\/p>\n","protected":false},"author":1,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"open","template":"","meta":{"footnotes":""},"class_list":["post-2","page","type-page","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/celso.world\/noograms\/wp-json\/wp\/v2\/pages\/2","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/celso.world\/noograms\/wp-json\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/celso.world\/noograms\/wp-json\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/celso.world\/noograms\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/celso.world\/noograms\/wp-json\/wp\/v2\/comments?post=2"}],"version-history":[{"count":14,"href":"https:\/\/celso.world\/noograms\/wp-json\/wp\/v2\/pages\/2\/revisions"}],"predecessor-version":[{"id":459,"href":"https:\/\/celso.world\/noograms\/wp-json\/wp\/v2\/pages\/2\/revisions\/459"}],"wp:attachment":[{"href":"https:\/\/celso.world\/noograms\/wp-json\/wp\/v2\/media?parent=2"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}