# Aplikacje LLM i agenci

EN: **LLM applications and agents** · id `applications`

Path: [AI Atlas](https://aiatlasgraph.com/llms.txt) › Aplikacje LLM i agenci

## Description

- PL: Budowanie systemów korzystających z modeli i narzędzi.
- EN: Building systems that use models and tools.

## In one sentence

- PL: To dziedzina budowania programów, w których sztuczna inteligencja czyta, rozmawia i korzysta z narzędzi.
- EN: This field builds programs that use artificial intelligence to read information, hold conversations and use tools.

## Problem it solves

- PL: Musisz ręcznie szukać informacji i przenosić je między programami, żeby załatwić nawet prostą sprawę.
- EN: Even a simple task can require you to find information and move it between several programs by hand.

## What it does

- PL: Łączy model językowy z dokumentami i narzędziami oraz ustala zasady, które kierują jego pracą.
- EN: It connects a language model to documents and tools, with rules that control which information and actions it can use.

## Everyday example

- PL: Prosisz firmowego asystenta o znalezienie warunków zwrotu i przygotowanie odpowiedzi dla klienta.
- EN: You ask a workplace assistant to find the returns policy and draft a reply to a customer.

## This concerns you if you…

- PL: chcesz załatwiać sprawy przez rozmowę z asystentem
- PL: przenosisz informacje między dokumentami a firmowymi programami
- EN: want to get things done by talking to an assistant
- EN: move information between documents and workplace software

## Keywords

RAG · function calling · prompt templates · context engineering · agentic workflows · LangGraph · LlamaIndex · MCP · JSON Schema · conversation state · human-in-the-loop · memory retrieval

## How it works

- PL: Aplikacja buduje kontekst z instrukcji, danych i historii, a następnie wywołuje model lub sekwencję modeli. Kod orkiestracji łączy generowanie z wyszukiwaniem, narzędziami, pamięcią i interfejsem oraz sprawdza wyniki przed ich użyciem.
- EN: The application assembles context from instructions, data and history, then calls a model or a sequence of models. Orchestration code connects generation with retrieval, tools, memory and the interface, and checks outputs before use.

## Limits

- PL: Jakość zależy od modelu, kontekstu i integracji. Opóźnienia oraz koszt rosną z liczbą wywołań, a zmiana modelu może wymagać ponownej ewaluacji całej aplikacji.
- EN: Quality depends on the model, context and integrations. Latency and cost grow with additional calls, while a model change may require reevaluating the entire application.

## Key terms

- PL: **kontekst**: Informacje dostępne modelowi podczas konkretnego wywołania.
- PL: **orkiestracja**: Kod ustalający kolejność wywołań i przepływ ich wyników.
- PL: **kontrakt wyjścia**: Uzgodniony format i warunki poprawności wyniku.
- EN: **context**: Information available to the model during a particular call.
- EN: **orchestration**: Code controlling call order and the flow of results.
- EN: **output contract**: The agreed format and validity requirements for an output.

## Know first

- [Transformery i attention (Transformers and attention)](https://aiatlasgraph.com/t/transformers.md)
- [Zbiory testowe i regresje (Evaluation datasets and regression tests)](https://aiatlasgraph.com/t/evalsets.md)

## Subtopics (11 direct, 76 in total)

- [Agenci i orkiestracja (Agents and orchestration)](https://aiatlasgraph.com/t/agentsnode.md): Action loops, delegation and control over execution.
  - [Agenci przy komputerze (Computer and browser use)](https://aiatlasgraph.com/t/computeruse.md): An agent clicking in a browser and on a desktop.
    - [Granica zaufania SA-CUA dla agentów komputerowych (SA-CUA trust boundary for computer-use agents)](https://aiatlasgraph.com/t/sacua.md): A dual-model architecture that masks the screen by default, reveals trusted regions using accessibility-tree rules, and separately endorses values from untrusted content.
    - [Syntetyczne środowiska komputerowe do treningu (Synthetic computer-use environments for agent training)](https://aiatlasgraph.com/t/softwaregym.md): Pipelines that automate synthesis, resetting and verification of executable training environments for computer-use agents.
  - [Agenci w tle (Ambient and scheduled agents)](https://aiatlasgraph.com/t/ambient.md): Agents that wake up on their own: on a schedule or when something changes.
  - [Autor i recenzent (Maker–checker review)](https://aiatlasgraph.com/t/makerchecker.md): No agent grades its own work; a second one assumes the output is broken and looks for where.
  - [Dyfuzyjne modele językowe w agentach badań głębokich (Diffusion language models in deep research agents)](https://aiatlasgraph.com/t/dllmdeepresearch.md): Deep-research systems in which a diffusion language model powers model-bound stages such as planning, decomposition and report synthesis.
    - [Wczesna ekstrakcja podzapytań ze stanów odszumiania (Early subquery extraction from diffusion denoising states)](https://aiatlasgraph.com/t/dfast.md): A decomposition mode that reads intermediate denoising states of a diffusion model and extracts usable subqueries before generation finishes.
  - [Hybrydowe agenty rozumujące (Hybrid reasoning agents)](https://aiatlasgraph.com/t/hybridagents.md): An agent dynamically switches between latent-space reasoning and text for tool calls and answers.
  - [Orkiestracja subagentów (Subagent orchestration)](https://aiatlasgraph.com/t/subagents.md): Deciding when to launch subagents and how to hand work to them.
    - [Rynkowa samoorkiestracja rojów agentów (Market-based post-training for decentralized agent swarms)](https://aiatlasgraph.com/t/mpt.md): A post-training framework in which decentralized agent swarms jointly learn action policies and self-orchestration through bids.
    - [Samonadzór wieloagentowy (MASS) (Multi-agent self-supervision (MASS))](https://aiatlasgraph.com/t/mass.md): A recursive self-improvement method in which one model alternates between improving multi-agent workflows and fine-tuning its own weights on selected trajectories.
  - [Planowanie przez wyobrażone przyszłe stany (Planning with imagined future observations)](https://aiatlasgraph.com/t/agima.md): A technique combining policy and world model that generates visual consequences of hypothetical actions and selects a plan by comparing future states.
  - [Protokoły agent–agent (Agent-to-agent protocols)](https://aiatlasgraph.com/t/a2a.md): An agreed way for agents from different systems to talk.
  - [Role i tożsamość agentów (Agent roles and identity)](https://aiatlasgraph.com/t/agentroles.md): A named agent with a scope and standing instructions instead of an anonymous session.
  - [Routing między modelami (Model and harness routing)](https://aiatlasgraph.com/t/routing.md): Matching the model and runtime to the kind of task and its cost.
  - [Samodzielna nawigacja drzewa stanów agenta (Agent self-navigation in checkpointed search trees)](https://aiatlasgraph.com/t/treeflex.md): A technique in which an agent saves intermediate states as a tree and chooses which node to resume, separating tree-level reflection from tool execution.
  - [Trwałe naprawy agentów przez adaptery lekcji (Persistent agent repair with lesson adapters)](https://aiatlasgraph.com/t/selfdirection.md): A mechanism in which an agent, after failure, distils a lesson from its trajectory into a lightweight LoRA adapter while keeping base weights frozen.
  - [Umiejętności agentów (Agent skills)](https://aiatlasgraph.com/t/agentskills.md): A catalog of folders with metadata, instructions and resources, loaded in layers according to the task match.
    - [Format pliku umiejętności (Skill file format)](https://aiatlasgraph.com/t/skillformat.md): A folder with an instruction file whose header has at least a name and description, matching the folder name.
    - [Skrypty w paczce umiejętności (Bundled skill scripts)](https://aiatlasgraph.com/t/skillscripts.md): Executable code bundled with a skill and run as a tool instead of being generated into context.
    - [Stopniowe wczytywanie umiejętności (Progressive disclosure of skills)](https://aiatlasgraph.com/t/skilldisclosure.md): Three-level skill loading: metadata at startup, body after selection, resources on demand.
  - [Uprząż agenta (Agent harness)](https://aiatlasgraph.com/t/agentharness.md): The loop, tools, context and constraints that turn a model call into an agent that carries out work.
    - [Budżety czasu dla agentów (Time-budgeted agent control)](https://aiatlasgraph.com/t/agenttimebudget.md): Harness and learning mechanisms that help an agent adhere to a wall-clock time budget.
    - [Kompozycyjne przeszukiwanie uprzęży (Compositional harness search)](https://aiatlasgraph.com/t/harnesscomposition.md): Modular harness with pools of implementations; a contextual bandit selects combinations for evaluation.
    - [Mapa dokumentów dla agenta (Repository map for agents)](https://aiatlasgraph.com/t/harnessdocs.md): The agent's entry point is a short map to versioned documents, not a full manual in one file.
    - [Mechaniczne reguły środowiska agenta (Mechanical invariants for agent environments)](https://aiatlasgraph.com/t/harnessinvariants.md): Architectural and style invariants enforced by tests and linters, with a repair hint in the error.
    - [Przekazanie pracy między sesjami (Cross-session handoff for agents)](https://aiatlasgraph.com/t/sessionhandoff.md): Passing state between context windows through a log, a feature list and change history, not through compaction alone.
    - [Uczenie uprzęży agentów (Harness learning)](https://aiatlasgraph.com/t/harnesslearning.md): Meta-learning over harness code: a proposer generates edits, reinforcement learning on execution outcomes.
      - [Adaptacja uprzęży podczas RL (Harness adaptation during agentic reinforcement learning)](https://aiatlasgraph.com/t/harnessadaptrl.md): RL in which the policy and harness adapt together based on training trajectories.
      - [Koewolucja modelu i uprzęży agenta (Weight-harness co-evolution)](https://aiatlasgraph.com/t/weightharnesscoevo.md): A single policy acts as proposer and reflector; harness edit logs train its weights.
      - [Rekurencyjna uprżąż z rewizją wsteczną (Recurrent harness with backward revision)](https://aiatlasgraph.com/t/reharness.md): A harness that adapts textual state by reading and writing reflections, then revising them from the outcomes of their downstream use.
      - [Skalowanie uprzęży w agenticznym RL (Harness-diversity scaling for agentic reinforcement learning)](https://aiatlasgraph.com/t/harnessscaling.md): Scaling agentic RL by training on diverse harness configurations instead of one fixed one.
    - [Zgodność intencji z wykonaniem narzędzia (Tool-call intent–execution correspondence)](https://aiatlasgraph.com/t/intentexecution.md): The property that the action actually executed by a tool matches the action denoted by the call under the tool contract.
  - [Wspólna przestrzeń agentów (Shared agent workspace)](https://aiatlasgraph.com/t/blackboard.md): A place where agents see each other’s work. Classic name: blackboard architecture.
  - [Wyuczone wczesne zatrzymywanie nieudanych uruchomień agentów (Learned early stopping of failing agent runs)](https://aiatlasgraph.com/t/stoploss.md): External small monitor model trained on completed trajectories and verifier outcomes that predicts failure and aborts the run.
- [Agenci kodujący (AI-assisted coding)](https://aiatlasgraph.com/t/aicoding.md): Writing, reviewing and maintaining code with agents.
  - [Kernele pisane przez LLM (LLM-generated kernels)](https://aiatlasgraph.com/t/llmkernels.md): Models and agents that write and speed up GPU code themselves — with RL and evolutionary search.
  - [Migracja i modernizacja kodu z AI (AI-assisted code migration and modernisation)](https://aiatlasgraph.com/t/codemigration.md): Models and agents rewrite large amounts of existing code: library upgrades, interface changes, translation between languages.
  - [Podpowiadanie kodu w edytorze (Inline code completion (fill-in-the-middle))](https://aiatlasgraph.com/t/codecompletion.md): A model writes the next lines or blocks of code at the cursor; systems trained with fill-in-the-middle can use code both before and after the gap.
  - [Programowanie od specyfikacji (Spec-driven development with coding agents)](https://aiatlasgraph.com/t/specdriven.md): A structured specification is created first, often together with a technical plan and task list; these artifacts guide the agent and may remain a reference point for later changes.
  - [Przegląd kodu przez AI (AI code review)](https://aiatlasgraph.com/t/aicodereview.md): A model analyses a code change (pull request) and leaves comments on bugs, risks and style before human review.
  - [Sterowanie przed pierwszą edycją kodu (Pre-edit planning intervention for coding agents)](https://aiatlasgraph.com/t/pause.md): A single planning intervention before the first source edit in coding agents.
  - [Vibe coding: tworzenie bez przeglądania kodu (Vibe coding (creating software without reviewing generated code))](https://aiatlasgraph.com/t/vibecoding.md): Iterative software creation through prompts and behavioral testing without reading or understanding the generated code; unlike responsible AI-assisted programming, the generated changes are not reviewed.
- [Asystenci i chatboty (Chatbots and assistants)](https://aiatlasgraph.com/t/chatbots.md): Conversation as the way a model is delivered.
- [Frameworki agentowe (Agent frameworks)](https://aiatlasgraph.com/t/frameworks.md): Ready-made skeletons for building agents and LLM applications.
- [Generatywne UI (Generative UI)](https://aiatlasgraph.com/t/genui.md): Interface built by the model as the conversation goes.
- [Narzędzia i integracje (Tool use and integration)](https://aiatlasgraph.com/t/tools.md): Connecting a model to functions and external systems.
  - [MCP (Model Context Protocol)](https://aiatlasgraph.com/t/mcp.md): A protocol for exposing tools and context to model-based applications.
  - [Structured outputs i walidacja (Structured outputs and validation)](https://aiatlasgraph.com/t/structured.md): Forcing the shape of an answer and checking it before use.
- [Pamięć i stan (Memory and state)](https://aiatlasgraph.com/t/memory.md): Keeping information between steps and conversations.
  - [Agentyczna pamięć z dynamicznymi powiązaniami (Agentic memory with dynamic linking)](https://aiatlasgraph.com/t/agenticmem.md): Technique for agentic memory organisation through dynamic note creation and linking inspired by the Zettelkasten method.
  - [Hierarchiczne zarządzanie kontekstem wirtualnym (Virtual context management with hierarchical memory)](https://aiatlasgraph.com/t/hierarchicalmem.md): Technique for managing virtual context via a memory hierarchy and agent interrupts, inspired by operating systems.
  - [Nietrwałe wskazówki w pamięci agentów (Anti-Persistence for agent memory)](https://aiatlasgraph.com/t/antipersistence.md): A memory design that accumulates trajectories and outcomes, synthesizes temporary guidance on demand, and discards it after execution.
  - [Pamięć proceduralna (Procedural memory)](https://aiatlasgraph.com/t/procedural.md): Saved recipes for solved problems, ready for reuse.
  - [Pętla samodoskonalenia (Reflection and feedback loops)](https://aiatlasgraph.com/t/reflection.md): Reasons for rejections are written back into instructions and memory, so the same mistake does not survive twice.
- [Prompty i kontekst (Prompting and context engineering)](https://aiatlasgraph.com/t/prompting.md): Choosing the instructions and information available at call time.
  - [Budowa polecenia: instrukcje, rola i znaczniki (Prompt structure: instructions, roles and delimiters)](https://aiatlasgraph.com/t/promptstructure.md): Laying out a prompt with clear instructions, purpose, a role in the system prompt, data sections separated by delimiters and an output format description.
  - [Programowa optymalizacja promptów (Programmatic prompt optimisation)](https://aiatlasgraph.com/t/promptopt.md): Automatically tuning prompts, and sometimes weights, against an evaluation score.
  - [Przykłady w poleceniu (few-shot) (Few-shot prompting and in-context learning)](https://aiatlasgraph.com/t/fewshotprompting.md): A few input–output pairs placed in the prompt show the model the task and answer format without changing its weights.
  - [Zarządzanie oknem kontekstu (Context window management: placement, compaction and context rot)](https://aiatlasgraph.com/t/contextmanagement.md): Deciding which tokens enter a limited context and in what order: placement, trimming, summarisation and external notes.
- [Przetwarzanie dokumentów (Document AI)](https://aiatlasgraph.com/t/docai.md): Pulling content and structure out of documents.
  - [Analiza układu i parsowanie dokumentów (Document layout analysis and parsing)](https://aiatlasgraph.com/t/layoutanalysis.md): Detecting page blocks (headings, paragraphs, tables, figures, footnotes), determining reading order and exporting structured text.
  - [Wyciąganie pól z formularzy i faktur (Key information extraction from forms and receipts)](https://aiatlasgraph.com/t/kie.md): Mapping document fragments to named fields (e.g. tax ID, gross amount, date) using text, its position and the page's appearance.
  - [Wyciąganie tabel z dokumentów (Table extraction and structure recognition)](https://aiatlasgraph.com/t/tableextraction.md): Detecting a table on a page, reconstructing its rows, columns, headers and merged cells, and saving it in a machine-readable format.
- [Workflow i udział człowieka (Workflows and human-in-the-loop)](https://aiatlasgraph.com/t/workflows.md): Combining predictable stages with model and human decisions.
  - [Architektura zdarzeniowa (Event-driven AI)](https://aiatlasgraph.com/t/eventdriven.md): Pipelines triggered by events rather than by a click.
  - [Łańcuchy promptów (Prompt chaining)](https://aiatlasgraph.com/t/promptchaining.md): A complex task is split into a fixed sequence of simpler model calls in which one step's output is the next step's input.
- [Wyszukiwanie i RAG (Retrieval-augmented generation)](https://aiatlasgraph.com/t/retrieval.md): Fetching knowledge and using it while answering.
  - [Bazy wektorowe (Vector databases)](https://aiatlasgraph.com/t/vectordb.md): Storing and searching by vectors.
  - [Dzielenie dokumentów (Document chunking)](https://aiatlasgraph.com/t/chunking.md): Cutting documents into indexable pieces, preferably by structure and with metadata.
  - [Hierarchiczna nawigacja dokumentów modelami decyzyjnymi (Hierarchical document navigation with decision models)](https://aiatlasgraph.com/t/decisionrag.md): Navigation through a corpus hierarchy in which a decision model evaluates branches jointly at each level without autoregressive text generation.
  - [Przekształcanie zapytań (Query transformation)](https://aiatlasgraph.com/t/queryrewrite.md): Rewriting the user's question before it reaches the search.
  - [Reranking i fuzja wyników (Reranking and result fusion)](https://aiatlasgraph.com/t/reranking.md): Re-ordering retrieved chunks before they reach the model, for relevance and diversity.
  - [Wyszukiwanie hybrydowe (Hybrid retrieval)](https://aiatlasgraph.com/t/hybridsearch.md): Combining vector similarity with keyword matching, sometimes several vectors per document.
  - [Zapytania do danych strukturalnych (Query construction for structured data)](https://aiatlasgraph.com/t/queryconstruction.md): Translating a question into a database query language when the answer lives in tables or a graph.

## Sources

- [Hugging Face · Agents Course](<https://huggingface.co/learn/agents-course/en/unit0/introduction>) `agents` · reference: Agenci, narzędzia, frameworki, agentic RAG i ewaluacja.
- [FSDL · LLM Bootcamp (2023, archiwum)](<https://fullstackdeeplearning.com/llm-bootcamp/spring-2023/>) `bootcamp` · reference: Aplikacje LLM, prompting, retrieval, UX i LLMOps.
- [DeepLearning.AI — course catalogue](<https://www.deeplearning.ai/courses>) `dlai` · course catalogue: 42 topic labels with counts, about 130 courses.

---

AI Atlas revision 122 (2026-10-30) · HTML: https://aiatlasgraph.com/t/applications · index: https://aiatlasgraph.com/llms.txt · everything: https://aiatlasgraph.com/llms-full.txt · updated 2026-10-30