Skip to content
awesome-applied-ai

Inside I · Context engineering

The full index

Every entry across the six layers in one place. Narrow by topic, layer, licence model, maturity, deployment or project health — topics stack, so picking two shows only the entries carrying both. Maturity and status are my readings of the evidence; every row links to the source.

Everything here is a library you import. For the binaries, apps, MCP servers and corpora that sit around the work, see the kit.

133 / 133
01

AGENTS.md

Tells coding agents how to build, test and style a repository

Read by Codex, Cursor, Copilot, Gemini CLI, Zed, Aider and goose. No ratified versioned spec yet.

OrchestrationAAIFopen sourcewidely deployed
02

Anthropic count_tokens

Free endpoint returning billing-accurate counts including system prompt and tools

CachingCommercial (free)commercialwidely deployedsaas
03

Anthropic prompt caching

Up to four cache_control breakpoints with 5-minute or 1-hour TTL, refreshed on read

Writes cost 1.25x (5-min) or 2.0x (1-hr) input. Break-even at roughly 0.28 and 1.11 reads.

CachingCommercialcommercialwidely deployedsaas
04

Arize Phoenix

OTel-native tracing and eval workflow that runs fully local

EvaluationElastic-2.0source availablewidely deployedboth
05

Azure AI Search

Hybrid BM25, vector and semantic reranking with integrated vectorization

RetrievalCommercialcommercialwidely deployedsaas
06

Bedrock Knowledge Bases

Managed ingestion and retrieval over OpenSearch, Aurora or Neptune

RetrievalCommercialcommercialwidely deployedsaas
07

BGE (BAAI)

Open-weight embeddings; bge-reranker-v2-m3 is the common self-host default

RetrievalMIT / Apache-2.0open sourcewidely deployedself-host
08

Braintrust

Eval-first workflow with scoring, prompt playground and a VPC data plane

Proprietary schema first-class; OTel is an export path, not the internal model.

EvaluationProprietarycommercialwidely deployedboth
09

CLAUDE.md

Hierarchical project context with imports and per-directory scoping

OrchestrationAnthropiccommercialwidely deployed
10

Cohere Embed v4 / Rerank 4

Multimodal embeddings with Matryoshka dims, plus a cross-encoder rerank API

RetrievalCommercialcommercialwidely deployedsaas
11

Datadog LLM Observability

LLM spans inside existing APM, correlated with infrastructure telemetry

Usually already on the approved-vendor list, which matters more than features.

EvaluationProprietarycommercialwidely deployedsaas
12

Elasticsearch

BM25 plus dense and sparse retrieval, RRF hybrid, retrievers API

Added AGPL-3.0 in 2024, restoring OSI-approved status.

RetrievalAGPL-3.0 / ELv2 / SSPLsource availablewidely deployedboth
13

FAISS

ANN index library rather than a database; baseline for offline search

RetrievalMITopen sourcewidely deployedself-host
14

Gemini explicit caching

Declared cache object with 60-minute default TTL and a large token minimum

Adds storage cost per token-hour, unlike the other two.

CachingCommercialcommercialwidely deployedsaas
15

Instructor

Pydantic validation plus a retry loop around any provider

OrchestrationMITopen sourcewidely deployedself-host
16

Langfuse

Tracing, prompt management, evals and datasets with the strongest self-host story

Emits OTel-native spans. Survives enterprise procurement.

EvaluationMIT core + EEopen corewidely deployedboth
17

LangGraph

Graph runtime with checkpointed state machines, durable execution and human-in-the-loop

OrchestrationMITopen sourcewidely deployedboth
18

LangSmith

Trace and debug LangChain and LangGraph agents, with datasets and online evals

Self-hosting is gated behind the Enterprise tier.

EvaluationProprietarycommercialwidely deployedboth
19

LiteLLM

Gateway normalizing token accounting and cache headers across providers

CachingMITopen sourcewidely deployedboth
20

LlamaParse

VLM-based PDF parsing to Markdown; best-in-class on messy documents

RetrievalCommercialcommercialwidely deployedsaas
21

MCP

Client/server protocol for exposing tools, resources and prompts to models

Donated by Anthropic to the Linux Foundation in December 2025, removing single-vendor risk. Tool poisoning remains an OWASP-catalogued attack class.

OrchestrationLinux Foundation AAIFopen sourcewidely deployed
22

Microsoft Presidio

NER and regex PII detection, redaction and anonymisation

Deterministic, which makes it one of the few components here that behaves predictably.

GuardrailsMITopen sourcewidely deployedself-host
23

Milvus / Zilliz

Distributed billion-scale vector database; an LF AI & Data project

RetrievalApache-2.0open corewidely deployedboth
24

MongoDB Atlas Vector Search

Vector index inside MongoDB with native embedding and rerank APIs

RetrievalCommercial / SSPL corecommercialwidely deployedsaas
25

OpenAI automatic caching

Zero-config prefix cache with a 1,024-token minimum and no write surcharge

CachingCommercialcommercialwidely deployedsaas
26

OpenSearch

Apache-2.0 fork with k-NN, RRF hybrid and neural sparse retrieval

RetrievalApache-2.0open sourcewidely deployedboth
27

pgvector

Postgres extension: HNSW/IVFFlat vector indexes in your existing database

RetrievalPostgreSQLopen sourcewidely deployedself-host
28

Pinecone

Serverless managed vector database with hybrid search and namespaces

RetrievalCommercialcommercialwidely deployedsaas
29

Qdrant

Rust vector database with payload filtering, quantization and on-disk HNSW

RetrievalApache-2.0open corewidely deployedboth
30

RAGAS

RAG-component metrics: faithfulness, context precision and recall, answer relevancy

Scores a single retrieve-then-generate turn. No trajectory primitives.

EvaluationApache-2.0open sourcewidely deployedself-host
31

SGLang RadixAttention

Radix tree over token sequences for longest-prefix KV match, on by default

CachingApache-2.0open sourcewidely deployedself-host
32

Sub-agent context isolation

Spawn agents with private windows so only summaries reach the parent

A pattern rather than a product, and often more effective than a compression library.

CachingPatternopen sourcewidely deployed
33

Tantivy

Rust full-text search library embedded in many higher-level tools

RetrievalMITopen sourcewidely deployedself-host
34

Temporal

General durable execution treating agents as replayable workflows

Infrastructure rather than an agent library. Often the right substrate underneath one.

OrchestrationMITopen corewidely deployedboth
35

tiktoken

OpenAI BPE tokenizer giving exact counts for OpenAI models only

CachingMITopen sourcewidely deployedself-host
36

Unstructured

Thirty-plus formats to normalized elements, with connectors and VPC deployment

RetrievalApache-2.0 coreopen corewidely deployedboth
37

Vertex AI Search

Managed retrieval and grounding with connectors and a ranking API

RetrievalCommercialcommercialwidely deployedsaas
38

Vespa

Search and ranking engine with tensor compute and ML ranking phases

RetrievalApache-2.0open corewidely deployedboth
39

vLLM prefix caching

Block-level content-hashed KV reuse, on by default in V1

CachingApache-2.0open sourcewidely deployedself-host
40

Weaviate

Native BM25 and dense hybrid search, multi-tenancy, module system

RetrievalBSD-3open corewidely deployedboth
41

XGrammar

Pushdown-automaton constrained decoding, the default in vLLM, SGLang and TRT-LLM

OrchestrationApache-2.0open sourcewidely deployedself-host
42

A2A

Peer agent discovery via Agent Cards, plus task delegation

OrchestrationLinux Foundationopen sourceproduction viable
43

Anthropic context editing

Server-side clearing of stale tool results and thinking blocks before token counting

MemoryCommercialcommercialproduction viablesaas
44

Anthropic memory tool

Claude issues CRUD against a /memories directory; you own the storage backend

Stores nothing itself. Budget for path-traversal defenses, per-tenant isolation and retention policy.

MemoryCommercialcommercialproduction viablesaas
45

Azure Prompt Shields

Managed moderation and prompt-injection filters

Published work reports character-injection evasion approaching total bypass.

GuardrailsProprietarycommercialproduction viablesaas
46

Chonkie

Chunking library covering token, semantic, recursive and late chunking

RetrievalMITopen sourceproduction viableself-host
47

Chroma

Embedded single-node vector store with a Rust core

RetrievalApache-2.0open coreproduction viableboth
48

Claude Agent SDK

The harness behind Claude Code: subagents, hooks, compaction, MCP-native

OrchestrationAnthropiccommercialproduction viableboth
49

ColBERT / PLAID

Late-interaction retrieval with token-level MaxSim; higher index cost

RetrievalApache-2.0open sourceproduction viableself-host
50

Comet Opik

Tracing, evaluation and guardrail hooks with no enterprise feature gating

EvaluationApache-2.0open sourceproduction viableboth
51

Credo AI

Risk registers, control mappings and Annex IV technical documentation

GuardrailsProprietarycommercialproduction viablesaas
52

DeepEval

Pytest-style LLM unit tests covering RAG, conversational and agent metrics

EvaluationApache-2.0open sourceproduction viableboth
53

Docling

IBM Research layout and table pipeline; local PDF-to-Markdown, no per-page cost

RetrievalMITopen sourceproduction viableself-host
54

Firecrawl

Crawls sites to clean Markdown or JSON and handles JS rendering

AGPL — check policy before embedding.

RetrievalAGPL-3.0 coreopen coreproduction viableboth
55

FlashRank

Small CPU-only reranker for latency-constrained pipelines

RetrievalApache-2.0open sourceproduction viableself-host
56

Gemini context caching

Explicit cached-content handles or implicit auto-cache on repeated prefixes

MemoryCommercialcommercialproduction viablesaas
57

Google ADK

Agent SDK with evaluation and a deployment path into Vertex Agent Engine

OrchestrationApache-2.0open sourceproduction viableboth
58

Graphiti

Temporal knowledge graph with edge validity intervals and hybrid graph-vector search

RetrievalApache-2.0open sourceproduction viableboth
59

Graphiti

Bi-temporal graph where edges carry valid-from and invalid-at, so facts expire

Temporal invalidation is the clearest differentiator against native memory tools.

MemoryApache-2.0open sourceproduction viableself-host
60

Guardrails AI

Output validators and schema enforcement via a validator hub

GuardrailsApache-2.0open sourceproduction viableself-host
61

Helicone

Proxy-based logging and caching with cache-hit-rate dashboards

Acquired by Mintlify in March 2026 and now in maintenance mode. Plan migration if you are on it.

stalledEvaluationApache-2.0open sourceproduction viableboth
62

Holistic AI

AI risk assessment, bias auditing and governance reporting

GuardrailsProprietarycommercialproduction viablesaas
63

HoneyHive

Tracing, evaluation and dataset curation with VPC deployment

EvaluationProprietarycommercialproduction viableboth
64

IBM watsonx.governance

Model inventory, risk workflows and compliance evidence inside the IBM stack

GuardrailsProprietarycommercialproduction viableboth
65

Inspect AI

UK AI Security Institute harness with solvers and scorers, strong for safety evals

EvaluationMITopen sourceproduction viableself-host
66

Jina AI

Open-weight embeddings and multimodal rerankers that score page images

RetrievalMixed OSS + APIopen coreproduction viableboth
67

Lakera Guard

Commercial injection and jailbreak detection API

GuardrailsProprietarycommercialproduction viablesaas
68

LanceDB

Serverless vector database on the Lance columnar format, reads from object storage

RetrievalApache-2.0open coreproduction viableboth
69

LangChain 1.x

Integration layer, now a thin facade over the LangGraph runtime

Third significant rewrite. Adopt LangGraph directly and treat integrations as optional glue.

OrchestrationMITopen sourceproduction viableself-host
70

Llama Guard / Prompt Guard

Hazard-taxonomy classifier plus a fast injection pre-filter

Defeated by prompt-overflow inputs that it catches in short context.

GuardrailsMeta communitysource availableproduction viableself-host
71

LlamaIndex

Ingestion, parsing, indexing and retrieval, with agent features added later

A retrieval library with agent features, not an orchestration spine.

OrchestrationMITopen sourceproduction viableself-host
72

llguidance

Low-latency grammar engine with the strongest measured JSON validity

OrchestrationMITopen sourceproduction viableself-host
73

LLM Guard

Scanner suite covering PII, toxicity and injection heuristics

GuardrailsMITopen sourceproduction viableself-host
74

LMCache

Cross-instance KV cache sharing and offload to CPU or NVMe

The answer for multi-node; SGLang alone suffices single-node.

CachingApache-2.0open sourceproduction viableself-host
75

Marker

PDF and EPUB to Markdown, GPU-accelerated and batch-oriented

RetrievalGPL-3.0 + commercialsource availableproduction viableself-host
76

Mastra

TypeScript-native agents, workflows, RAG and evals

OrchestrationApache-2.0open sourceproduction viableboth
77

Meilisearch

Lightweight full-text and hybrid search for in-application search

RetrievalMITopen coreproduction viableboth
78

Mem0

Extracts facts into a vector, graph and KV store scoped by user, session or agent

MemoryApache-2.0open coreproduction viableboth
79

Microsoft Agent Framework

AutoGen and Semantic Kernel merged; .NET and Python agents plus workflow graphs

OrchestrationMITopen sourceproduction viableboth
80

Microsoft GraphRAG

Entity extraction, community detection and hierarchical summaries

Indexing cost is substantial at corpus scale.

RetrievalMITopen sourceproduction viableself-host
81

Mixedbread

Open-weight embedding and reranker models plus a hosted API

RetrievalApache-2.0open coreproduction viableboth
82

MLflow Tracing

GenAI tracing inside an existing MLflow registry estate

EvaluationApache-2.0open sourceproduction viableboth
83

NeMo Guardrails

Colang dialog-flow rails and topic control that orchestrate other classifiers

GuardrailsApache-2.0open sourceproduction viableself-host
84

neo4j-graphrag-python

Official Neo4j retrievers and knowledge-graph construction pipeline

RetrievalApache-2.0open sourceproduction viableself-host
85

OpenAI Agents SDK

Minimal agent loop with handoffs, guardrails, tracing and sessions

OrchestrationMITopen sourceproduction viableboth
86

OpenAI Evals

Registry-based eval templates, OpenAI-centric

EvaluationMITopen sourceproduction viableself-host
87

OpenAI Responses API state

Server-stored response items chained via previous_response_id, 30-day retention

MemoryCommercialcommercialproduction viablesaas
88

Outlines

FSM-based structured generation, largely displaced by XGrammar

OrchestrationApache-2.0open sourceproduction viableself-host
89

ParadeDB

Postgres extension adding Tantivy-backed BM25 for in-database hybrid search

AGPL — check policy before embedding.

RetrievalAGPL-3.0 + commercialopen coreproduction viableself-host
90

pgvectorscale

StreamingDiskANN index layered on pgvector for larger-than-memory workloads

RetrievalPostgreSQLopen sourceproduction viableself-host
91

promptfoo

Declarative YAML eval matrices plus automated red-teaming

EvaluationMITopen sourceproduction viableboth
92

Pydantic AI

Type-safe agent layer with schema validation and dependency injection first-class

OrchestrationMITopen sourceproduction viableself-host
93

Qwen3-Embedding / Reranker

Open-weight embedding and reranker models topping MTEB in 2026

RetrievalApache-2.0open sourceproduction viableself-host
94

Reducto

Agentic OCR-correction parsing with on-prem, SOC 2 Type II and HIPAA

RetrievalCommercialcommercialproduction viableboth
95

Turbopuffer

Object-storage-native search with SPFresh index, BM25 and sparse vectors

RetrievalCommercialcommercialproduction viablesaas
96

Typesense

Typo-tolerant keyword search with vector fields and a simple ops model

RetrievalGPL-3.0open coreproduction viableboth
97

Vectara

Managed RAG with HHEM hallucination scoring; enterprise-only since 2026

Self-service tier discontinued in 2026.

RetrievalCommercialcommercialproduction viablesaas
98

VectorChord

Postgres extension with disk-friendly IVF/RaBitQ; successor to pgvecto.rs

AGPL — check policy before embedding.

RetrievalAGPL-3.0 + commercialopen coreproduction viableself-host
99

Voyage AI

Domain-tuned embeddings and rerankers for legal, code and finance

Acquired by MongoDB, February 2025.

RetrievalCommercialcommercialproduction viablesaas
100

W&B Weave

Trace and eval layer atop Weights & Biases

EvaluationApache-2.0 SDKopen coreproduction viableboth
101

Zep

Async summarization and entity extraction layered over Graphiti

MemoryApache-2.0open coreproduction viableboth
102

AGNTCY

Directory, identity, messaging and observability for agent meshes

OrchestrationLinux Foundationopen sourceearly
103

Ax

TypeScript equivalent of DSPy's signature-driven programs

OrchestrationApache-2.0open sourceearlyself-host
104

Chunkr

Parsing and semantic chunking API, self-hostable

AGPL — check policy before embedding.

RetrievalAGPL-3.0open coreearlyboth
105

Cognee

Pipeline turning documents into a queryable memory graph

RetrievalApache-2.0open sourceearlyself-host
106

Cognee

ECL pipeline building an ontology-typed graph alongside a vector index

MemoryApache-2.0open sourceearlyself-host
107

DBOS

Durable execution using Postgres as the workflow source of truth

OrchestrationMITopen coreearlyboth
108

GEPA

Reflective evolutionary prompt optimization, shipping inside DSPy

OrchestrationMITopen sourceearlyself-host
109

Hindsight

Retain, recall and reflect over memory banks with fact extraction and an MCP server

MemoryMITopen sourceearlyself-host
110

LangMem

Extracts and updates memories into a LangGraph store

Recent activity is largely dependency bumps, with no tagged releases.

slowingMemoryMITopen sourceearlyself-host
111

Letta (ex-MemGPT)

LLM-as-OS paging between main context, recall and archival stores

Core repo took ~5 commits in three months; development moved to sibling repos. Still widely recommended in roundups.

slowingMemoryApache-2.0open coreearlyboth
112

LightRAG

Flat dual-level graph index, far cheaper to build than GraphRAG

RetrievalMITopen sourceearlyself-host
113

llms.txt

Site-level markdown index intended for LLM consumers

No major AI platform has confirmed reading it. Optional SEO hygiene, not context engineering.

OrchestrationProposalopen sourceearly
114

MemOS

OS-style abstraction over plaintext, activation (KV) and parametric memory

MemoryApache-2.0open sourceearlyself-host
115

OTel GenAI conventions

Semantic conventions for gen_ai.* spans, metrics and events

Every document in the spec repo is still marked Development as of July 2026. Emit OTel anyway as a lock-in hedge.

EvaluationApache-2.0open sourceearly
116

Ragie

Managed ingestion and retrieval API with connectors and citations

RetrievalCommercialcommercialearlysaas
117

Supermemory

Memory API and RAG over user context, with a router SDK and MCP

MemoryMIT clientopen coreearlyboth
118

Trustible

Regulatory mapping and AI use-case registry for compliance teams

GuardrailsProprietarycommercialearlysaas
119

ZeroEntropy

zerank reranker models at the top of several 2026 leaderboards

RetrievalOSS weights + APIopen coreearlyboth
120

AutoGen

Research multi-agent conversation framework, superseded by Agent Framework

Maintenance mode. Greenfield work here is building on a dead branch.

deprecatedOrchestrationMITopen sourceresearchself-host
121

DSPy

Declarative signatures and modules that compile prompts against a metric

Offline optimizers, Python-only, no gateway or observability story.

OrchestrationMITopen sourceresearchself-host
122

LLMLingua

Token-classification prompt compressor achieving 2-5x reduction

Last release April 2024, no commits since April 2026. Still recommended in roundups.

stalledCachingMITopen sourceresearchself-host
123

Semantic Kernel

.NET-first agent and plugin SDK, folded into Agent Framework

Maintenance mode since Agent Framework GA in April 2026.

deprecatedOrchestrationMITopen sourceresearchself-host
124

TextGrad

Backpropagates natural-language gradients through LLM pipelines

A paper you can import.

OrchestrationMITopen sourceresearchself-host
125

CrewAI

Role-based multi-agent crews with LLM-driven task routing

LLM-driven routing makes cost and latency non-deterministic and failures hard to reproduce. Widely downloaded, widely abandoned at the pilot boundary.

OrchestrationMITopen corefragileboth
126

Context Rot (Chroma)

Eighteen models degrade non-uniformly as input grows; includes a replication toolkit

CachingResearchopen source
127

Digital Omnibus on AI

First AI Act amendment; Regulation (EU) 2026/1744, in force 27 July 2026

Deferred Annex III high-risk obligations to 2 Dec 2027 and Annex I to 2 Aug 2028. GPAI enforcement on 2 Aug 2026 was NOT deferred.

GuardrailsRegulationcommercial
128

EU AI Act

Risk-tiered obligations with extraterritorial reach; Regulation (EU) 2024/1689

In force 1 Aug 2024. Prohibitions and AI literacy live 2 Feb 2025. GPAI obligations live 2 Aug 2025.

GuardrailsRegulationcommercial
129

ISO/IEC 42001:2023

Certifiable AI management system standard

The only auditable certificate in this space. Takes 6-12 months and increasingly appears in due-diligence questionnaires.

GuardrailsStandardcommercial
130

LongFuncEval

Long-context degradation measured specifically for tool and function calling

CachingResearchopen source
131

Lost in the Middle

U-shaped positional accuracy, with over 30% drop for mid-context evidence

CachingResearchopen source
132

NIST AI RMF 1.0

Voluntary US risk framework with a Generative AI profile

The cheapest credible starting point, and it maps onto ISO 42001. Roughly 3-6 months.

GuardrailsPublicopen source
133

NoLiMa

Needle-in-haystack without lexical overlap; scores collapse where standard NIAH saturates

CachingResearchopen source