1~
2Oi AI agents echoun metakinithei apo ereynitikes epideixes se systimata paragogis. Perissoteroi apo to 60% ton epicheirisiakon AI efarmogon anamenetai na perilamvanoun agentika stoicheia mechri to 2026. Alla i kataskevi agents apo to miden  -  diacheirisi vrogchon ergaleion, katastasis, mnimis, diacheirisis sfalmaton kai syntonismou pollalon agents  -  einai polyploki. Edoo mpainoun ta frameworks.
3~
4Tessera frameworks kyrarchoun to 2026: **LangGraph**, **CrewAI**, **OpenAI Agents SDK** kai **Claude Agent SDK**. Kathena proseggizei to idio provlima me themeliodoos diaforetiko tropo: na dosei sta LLM tin ikanotita na skeftontai, na schedkazoun, na chrisimopoioun ergaleia kai na synergazontai.
5~
6## Me mia matia
7~
8| Opseis | LangGraph | CrewAI | OpenAI Agents SDK | Claude Agent SDK |
9|--------|-----------|--------|-------------------|-----------------|
10| **Dimioyrgos** | LangChain | CrewAI Inc. | OpenAI | Anthropic |
11| **Architektoniki** | Vasismeni se grafous | Vasismeni se rolous | Vasismeni se paradoseis | Aftonomi vrogchos |
12| **Filosofia** | Megistos elegchos | Omadiki synergasia | Elachisti afairesi | Dose ston agent enan ypologisti |
13| **Glosses** | Python, TypeScript | Python | Python | Python, TypeScript |
14| **Ypostirixi montelon** | Opoiodipote (OpenAI, Claude, topiko) | Opoiodipote | Opoiodipote (para to onoma) | Mono Claude |
15| **GitHub stars** | ~29k | ~40k | ~21k | ~6k |
16| **Kalytero gia** | Polyploka stateful workflows | Ekseidikeysi pollalon agents | Routing kai triaz | Kodika kai ergasies me archeia |
17~
18## LangGraph: O kataskevastis grafon
19~
20To LangGraph montelopoiei ta workflows ton agents os **kateythynomena kyklika grafa**. Orizete komvous (synartiseis pou ekteloun ergasia) kai akmes (metavaseis metaxy tous, proairetika me synthikes). I katastasi reei mesa apo to grafo kai diatireitai meso checkpointing.
21~
22Afto einai to pio safes kai elegchomeno framework  -  syndeete kathe vima monoi sas.
23~
24```mermaid
25graph LR
26    Start --> Router[Router Node]
27    Router -->|needs research| Research[Research Node]
28    Router -->|needs code| Code[Code Node]
29    Research --> Synthesize[Synthesize Node]
30    Code --> Synthesize
31    Synthesize --> End
32```
33~
34### Vasikes ennoies
35~
36- **StateGraph**: o orismos tou grafou me typopoiimeni katastasi
37- **Nodes**: synartiseis Python pou metaschimatizoun tin katastasi
38- **Edges**: syndesmoi metaxy komvon, mporoun na einai me synthiki
39- **Checkpointing**: ensoomatomeni monimotita gia makrochrones roes ergasias
40~
41### Paradeigma kodika
42~
43```python
44from langgraph.graph import StateGraph, MessagesState, START, END
45from langchain_openai import ChatOpenAI
46~
47llm = ChatOpenAI(model="gpt-4o")
48~
49def call_agent(state: MessagesState):
50    response = llm.invoke(state["messages"])
51    return {"messages": [response]}
52~
53def should_continue(state: MessagesState):
54    last = state["messages"][-1]
55    if last.tool_calls:
56        return "tools"
57    return END
58~
59def call_tools(state: MessagesState):
60    # Execute tool calls and return results
61    results = []
62    for tool_call in state["messages"][-1].tool_calls:
63        result = execute_tool(tool_call)
64        results.append(result)
65    return {"messages": results}
66~
67graph = StateGraph(MessagesState)
68graph.add_node("agent", call_agent)
69graph.add_node("tools", call_tools)
70graph.add_edge(START, "agent")
71graph.add_conditional_edges("agent", should_continue, {"tools": "tools", END: END})
72graph.add_edge("tools", "agent")
73~
74app = graph.compile()
75result = app.invoke({"messages": [{"role": "user", "content": "What's the weather?"}]})
76```
77~
78### Dynata simeia
79~
80- Leptoereis elegchos kathe vimatos kai metavasis
81- Ensoomatomeno checkpointing kai human-in-the-loop
82- Pliri isotimia me TypeScript
83- Leitoyrgei me opoiondipote parocho LLM
84- Kalytero gia polyplokes roes ergasias me synthiki diaklandosi kai vrogchous
85~
86### Adynamies
87~
88- Apotomi kampyli mathisis  -  prepei na katanoite ennoies theorias grafon
89- Analytiko gia aples periptoseis  -  enas vasikos agent apaite perissotero boilerplate apo alla frameworks
90- I aposfalmotosi roon grafon mporei na einai dyskoli choris to LangSmith
91~
92### Timologisi
93~
94Open-source (MIT). To LangSmith (diacheirizomeni platforma paratirisimotitas) echei epiploches vathmides gia paragoigki parakolouthisi.
95~
96## CrewAI: O dimiourgos omadas
97~
98To CrewAI chrisimopoiei mia anthropini metafora: synthetete ena **pliroma** apo ekseidikeymenous agents, kathenas me enan **rolo**, enan **stocho** kai mia **istoria**. Oi agents synergaaontai se **ergasies** chrisimopoiontas **ergaleia**, syntonismenoi apo mia **diadikasia** (seiariaki, ierarchiki i synthesigmatiki).
99~
100Skeftite to san na prosthenete mia omada opou kathe melos echei sygekkrimeno titlo kai eidikotita.
101~
102```mermaid
103graph TD
104    Crew[Crew Manager] --> R[Researcher\nRole: Find data\nTools: WebSearch]
105    Crew --> W[Writer\nRole: Write content\nTools: FileWrite]
106    Crew --> E[Editor\nRole: Review quality\nTools: FileRead]
107    R --> Task1[Research Task]
108    W --> Task2[Writing Task]
109    E --> Task3[Review Task]
110    Task1 --> Task2 --> Task3
111```
112~
113### Vasikes ennoies
114~
115- **Agent**: mia persona me rolo, stocho, istoria kai ergaleia
116- **Task**: mia anathesi me perigrafi, anammenomeno apotelesma kai anatetheimeno agent
117- **Crew**: mia omada agent pou ergaazontai mazi
118- **Process**: stratigiki ektelesis (seiariaki, ierarchiki, synthesigmatiki)
119- **Flow**: epipedo orchestration vasismeno se gegonota gia ti syndesi pollalon pliromaton
120~
121### Paradeigma kodika
122~
123```python
124from crewai import Agent, Task, Crew, Process
125~
126researcher = Agent(
127    role="Senior Research Analyst",
128    goal="Find comprehensive data about the given topic",
129    backstory="You have 10 years of experience in technology research. "
130              "You are thorough and always verify facts from multiple sources.",
131    tools=[web_search_tool],
132    verbose=True,
133)
134~
135writer = Agent(
136    role="Technical Writer",
137    goal="Create clear, engaging technical content",
138    backstory="You write for a developer audience. "
139              "Your articles are practical and include code examples.",
140    tools=[file_tool],
141    verbose=True,
142)
143~
144research_task = Task(
145    description="Research the latest developments in WebAssembly in 2026. "
146                "Focus on WASI, Component Model, and production use cases.",
147    expected_output="A structured research document with key findings and sources.",
148    agent=researcher,
149)
150~
151writing_task = Task(
152    description="Write a blog post based on the research. "
153                "Include code examples and Mermaid diagrams.",
154    expected_output="A complete blog post in Markdown format.",
155    agent=writer,
156    context=[research_task],  # Writer receives researcher's output
157)
158~
159crew = Crew(
160    agents=[researcher, writer],
161    tasks=[research_task, writing_task],
162    process=Process.sequential,
163    verbose=True,
164)
165~
166result = crew.kickoff()
167print(result.raw)
168```
169~
170### Dynata simeia
171~
172- Diaisthitiki afairesi vasismeni se rolous  -  eykoli stin katanoisi
173- Pano apo 100 ensoomatomenees entaxis ergaleion
174- Koinochroni mnimi metaxy ton agents (vrachyprothesmi, makroprthesmi, onotitas)
175- Megaliteri koinotita (~40k GitHub stars)
176- Ierarchiki diadikasia me enan agent "diacheiristi" pou anathete kai epikyroni
177~
178### Adynamies
179~
180- Ligotero leptoereis elegchos apo to LangGraph  -  orizete rolous, ochi akriveis diadirmes ektelesis
181- I ierarchiki diadikasia mporei na einai aprobleepti otan oi agents diafounoun
182- I aposfalmotosi synomilion pollalon agents einai dyskoloteri apo roes enos agent
183~
184### Timologisi
185~
186Open-source pyriinas (dorean). CrewAI Platform: 99 $/mina (Teams) eos 120 chil. $/etos (Enterprise). I timologisi vasizetai se energa pliromata kai miniaia ekteleseis.
187~
188## OpenAI Agents SDK: O router
189~
190To OpenAI Agents SDK (pnevmatikos diadochos tou Swarm) epikentroonetai stis **paradoseis**  -  oi agents metaferoun synomilies se allous ekseidikeymenous agents. Einai to pio minimalistiko framework: agents, ergaleia, paradoseis kai guardrails. Aftoo einai ola.
191~
192```mermaid
193graph LR
194    User --> Triage[Triage Agent]
195    Triage -->|billing question| Billing[Billing Agent]
196    Triage -->|refund request| Refund[Refund Agent]
197    Triage -->|technical issue| Support[Support Agent]
198    Billing --> Response[Response]
199    Refund --> Response
200    Support --> Response
201```
202~
203### Vasikes ennoies
204~
205- **Agent**: modelo + odigies + ergaleia + paradoseis
206- **Handoff**: mia metafora se allon agent (montelopoiimeni os ergaleio pou to LLM mporei na kalesei)
207- **Guardrail**: epikyrosi eisodou/exodou pou ekteleitai parallila me ton agent
208- **Runner**: ektelei ti vrogcho tou agent
209- **Tracing**: ensoomatommeni paratirisimotita gia oles tis kliseis LLM, chrisis ergaleion kai paradoseis
210~
211### Paradeigma kodika
212~
213```python
214from agents import Agent, Runner, handoff, InputGuardrail, GuardrailFunctionOutput
215from pydantic import BaseModel
216~
217class SafetyCheck(BaseModel):
218    is_safe: bool
219    reason: str
220~
221async def content_safety(ctx, agent, input_text):
222    result = await Runner.run(
223        Agent(name="Safety", instructions="Check if input is safe. No PII."),
224        input_text,
225        context=ctx,
226    )
227    output = SafetyCheck.model_validate_json(result.final_output)
228    return GuardrailFunctionOutput(
229        output_info=output, tripwire_triggered=not output.is_safe
230    )
231~
232billing_agent = Agent(
233    name="Billing Agent",
234    instructions="You handle billing inquiries. Be precise with numbers.",
235    tools=[lookup_invoice, process_payment],
236)
237~
238refund_agent = Agent(
239    name="Refund Agent",
240    instructions="You process refund requests. Always verify the order first.",
241    tools=[lookup_order, issue_refund],
242)
243~
244triage_agent = Agent(
245    name="Triage Agent",
246    instructions="Route the customer to the right specialist. "
247                 "Ask clarifying questions if needed.",
248    handoffs=[billing_agent, refund_agent],
249    input_guardrails=[InputGuardrail(guardrail_function=content_safety)],
250)
251~
252result = await Runner.run(triage_agent, "I need a refund for order #4521")
253print(result.final_output)
254# The triage agent routes to refund_agent, which processes the refund
255```
256~
257### Dynata simeia
258~
259- Katharo montelo paradosis  -  fysiko gia workflows routing/triaz
260- Ta guardrails ektelountai parallila me tin ektelesi (fail-fast, mi-mplokariistika)
261- Ensoomatommeno panel tracing gia aposfalmotosi
262- Para to onoma, ypostiri modela ektos OpenAI
263- Elachisti afairesi  -  eykolo na katanoithei kai na epektathei
264~
265### Adynamies
266~
267- Ligotero orimi diacheirisi katastasis apo to LangGraph
268- Kamia ensoomatommeni monimotita i checkpointing
269- To oikosystima ergaleion triton einai mikrotero
270- I schedkiasi me epikentro tis paradoseis mporei na min tairiazei se kathe architektoniki
271~
272### Timologisi
273~
274Open-source (MIT). Plirnete ana token gia opoiodipote modelo chrisimooieitee.
275~
276## Claude Agent SDK: O programmatistis
277~
278To Claude Agent SDK akolouthei mia diaforetiki proseggisi: anti na orizete workflows i rolous, dinete ston agent ena **set ergaleion kai ton afinete na vrei monos tou pos na oloklirosi tin ergasia**. Chrisimopoiei tin idia aftonomi vrogcho pou trofodioti to Claude Code  -  diavase, drase, epiviose, epanalav.
279~
280```mermaid
281graph TD
282    Prompt[User Prompt] --> Loop[Autonomous Agent Loop]
283    Loop --> Reason[Reason about next step]
284    Reason --> Act[Execute tool]
285    Act --> Verify[Check result]
286    Verify -->|not done| Loop
287    Verify -->|done| Output[Final output]
288```
289~
290### Vasikes ennoies
291~
292- **query()**: to kyrio simeio eisodou pou xekina ti vrogcho tou agent
293- **Ensoomatomena ergaleia**: Read, Write, Edit, Bash, Glob, Grep, WebSearch, WebFetch
294- **Prosarmosmena ergaleia meso MCP**: orismos ergaleion os in-process MCP servers
295- **Sub-agents**: ekseidikeymenoi agents stous opoious o gonikos mporei na anathesi
296- **Sessions**: diatirisi plaissiou se polles allilepidrasis
297~
298### Paradeigma kodika
299~
300```typescript
301import { tool, createSdkMcpServer, query } from "@anthropic-ai/claude-agent-sdk";
302import { z } from "zod";
303~
304const searchDocs = tool(
305  "search_docs",
306  "Search the internal documentation for relevant information",
307  { query: z.string().describe("Search query") },
308  async ({ query }) => {
309    const results = await vectorStore.similaritySearch(query, 5);
310    return {
311      content: [{ type: "text", text: results.map(r => r.pageContent).join("\n\n") }],
312    };
313  }
314);
315~
316const docsServer = createSdkMcpServer({
317  name: "docs",
318  version: "1.0.0",
319  tools: [searchDocs],
320});
321~
322for await (const message of query({
323  prompt: "Find how authentication works in our system and write a summary",
324  options: {
325    mcpServers: { docs: docsServer },
326    allowedTools: ["Read", "Glob", "Grep", "mcp__docs__search_docs"],
327  },
328})) {
329  if (message.type === "result" && message.subtype === "success") {
330    console.log(message.result);
331  }
332}
333```
334~
335### Dynata simeia
336~
337- Protis taxis entaxi MCP  -  syndethite me opoiodipote oikosystima MCP servers
338- Ensoomatomena ergaleia gia leitourgies archion, terminal kai prosvasi sto diadiktyo
339- Aftomati sympyknosi plaisiou gia megales vaseis kodika
340- Paralliloi sub-agents gia polyplokes ergasies
341- Idia michani me to Claude Code  -  dokimasmeni se pragmatikes roes ergasias anaptixi
342~
343### Adynamies
344~
345- Mono modela Claude  -  kamia ypostirixi pollon parochon
346- Neotero framework me mikroteri koinotita
347- Apaite periballon Node.js akomi kai gia to Python SDK
348- Ligoteros safis elegchos workflow se sygkrisi me to LangGraph
349~
350### Timologisi
351~
352Open-source. Kanoniki timi token Claude API. Managed Agents (filoxenoumeni ekdosi): 0,08 $ ana ora synedriias prostheta apo ta kosti token.
353~
354## Pote na epilexete poio
355~
356```mermaid
357graph TD
358    Start{What's your priority?}
359    Start -->|Full control over workflow| LG[LangGraph]
360    Start -->|Multi-agent collaboration| CA[CrewAI]
361    Start -->|Routing and triage| OA[OpenAI Agents SDK]
362    Start -->|Coding and file automation| CS[Claude Agent SDK]
363~
364    LG --> LGU[Complex stateful workflows\nConditional branching\nHuman-in-the-loop]
365    CA --> CAU[Team of specialized agents\nResearch + writing pipelines\nContent generation]
366    OA --> OAU[Customer service routing\nMulti-step handoffs\nInput validation]
367    CS --> CSU[Code generation and review\nFile-heavy automation\nMCP tool ecosystem]
368```
369~
370### Epilexte LangGraph an:
371- Chriazeste akrivii elegcho se kathe vima tis rois ergasias
372- I periptosi chrisis sas perilamvanei polyploki synthiki logiki kai vrogchous
373- Thelete ensoomatommeni monimotita kai checkpoints human-in-the-loop
374- Chriazeste na chrisimopoiisete pollous parochous LLM stin idia roi ergasias
375~
376### Epilexte CrewAI an:
377- Thelete mia diaisthitiki afairesi vasismeni se rolous
378- I ergasia sas perilamvanei pollous agents me diaforetikes eidikotites
379- Chriazeste oi agents na synergaaontai kai na pernnoun plaisio metaxy tous
380- Ektimate ti megaliteri koinotita kai tis perisoteres ensoomatommenes entaxis
381~
382### Epilexte OpenAI Agents SDK an:
383- To kyrio montelo sas einai i drromologisi synomilion se eidikous
384- Chriazeste guardrails pou epikyrounoun eisodo/exodo parallila
385- Thelete tin aploysteni dynati afairesi me elachisto boilerplate
386- To ensoomatommeno tracing kai i paratirisimotita einai simantika
387~
388### Epilexte Claude Agent SDK an:
389- Oi agents sas prepei na diavazoun, grafoun kai ekteloun kodika
390- Thelete protis taxis entaxi me MCP servers
391- Chriazeste aftonomous agents pou epanalamvanoun kai aftodiiorthonontai
392- Idi chrisimopoieite to Claude kai thelete tin vathyteri entaxi
393~
394## Mporun na syndyastoun ta frameworks?
395~
396Nai. Ena synitho montelo einai i chrisi enos framework gia tin orchestration kai enos allou gia tos monadikous agents:
397~
398- **LangGraph** gia to synoliko grafo rois ergasias
399- **CrewAI** gia ena sykekkrimeno komvo pou apaite synergasia pollalon agents
400- **Claude Agent SDK** gia ypo-ergasies schetikes me kodika meso MCP
401- **OpenAI Agents SDK** gia triaz kai routing pros tous pelates
402~
403Ta frameworks den einai amoimaia apokleiomena. Chrisimopoiiste oti tairiazei se kathe meros tou systimatos sas.
404~
405## Symperasma
406~
407Kathe framework kanei mia safii epilogi:
408~
409- **LangGraph** veltiopolitei gia elegcho  -  eseis apofasizete gia kathe metavasi
410- **CrewAI** veltiopolitei gia synergasia  -  oi agents ergaazontai os omada
411- **OpenAI Agents SDK** veltiopolitei gia aplotita  -  elachisti afairesi, kathares paradoseis
412- **Claude Agent SDK** veltiopolitei gia aftonomia  -  doste tou ergaleia kai afiste ton na douleepsi
413~
414I sosti epiloyi exartatai apo ti roi ergasias sas, tin omada sas kai to yparcho stack sas. Dialexte afto pou tairiazei stin kyria periptosi chrisis sas, mathete to kala kai travaaxte ta alla otan ftasete sto dynato tous simeio.
415~