[::] Agentic Stacks
The live index for agent builders

What's in your
AI agent stack?

A neutral directory of the models, frameworks, memory, retrieval, eval, and MCP servers teams use to ship production agents — every entry tagged with its source, and no fabricated metrics.

Browse stacks Explore tools →
179 tools indexed
31 stacks
15 use cases
Why you can trust it

No fabricated metrics.

Most directories pad themselves with invented eval scores and made-up cost numbers. We don't. Every entry is tagged with where it came from — and real numbers only show up when real teams submit them.

How we source data →
public_docs

Sourced from vendor docs and official READMEs.

researched

Real stacks reconstructed from public talks, posts, and OSS code — cited.

reference

Synthesized reference architectures, not attributed to any real company.

community

Submitted by a real team running a real production stack.

Featured stacks

All stacks →

Public: Aider (OSS pair programmer reference)

researched
coding-assistant
Aider is fully open source, so its "stack" is well documented. It's a

Public: Claude Code as an agent harness

researched
coding-assistant
Claude Code itself is a documented agent harness with a public

Public: Continue.dev (OSS IDE coding assistant)

researched
coding-assistant
Continue.dev is an OSS coding assistant that runs as VS Code and

Public: Glean Assistant enterprise RAG stack

researched
rag-over-docs
Per their engineering blog post on building an enterprise AI assistant,

Public: Harvey legal AI platform

researched
legal-review
Harvey runs a multi-provider, auto-routed model stack: an OpenAI partnership

Public: incident.io AI SRE

researched
ops-incident-response
incident.io's AI SRE is a multi-agent investigation engine that auto-activates

By use case

Coding Assistant / Pair Programmer
Autonomous or semi-autonomous code generation, review, refactor.
6
RAG Over Internal Docs
Answer questions over private documentation, wikis, code.
4
Deep Research
Multi-step research with synthesis across many sources.
3
Data Analysis / BI
Query databases in natural language, generate reports.
3
Legal Document Review
Contract review, clause extraction, redlining.
2
Ops / Incident Response
Triage alerts, gather context, propose remediation.
2
Customer Support Triage
Auto-categorize, draft replies, escalate inbound support tickets.
2
Recruiting Sourcing
Source candidates, draft outreach, screen responses.
2

By component

MCP Servers
Model Context Protocol servers exposing tools/resources.
50
Model
Underlying LLM(s); primary + fallback + routing.
26
Agent Framework
SDK / library structuring agent loops, tool calls, state.
16
Harness / IDE
Runtime surface (Claude Code, Cursor, custom backend).
14
Retrieval / Vector Store
Embedding store, hybrid search, RAG infrastructure.
11
Voice Infrastructure
STT, TTS, telephony for voice agents.
10
Observability / Tracing
Run traces, prompt logging, cost tracking.
9
Eval
Test harness, golden sets, regression detection.
7

Be first to share your stack

Community submissions open soon. Get the digest and we'll tell you the moment they do.

Get the weekly digest

New stacks, new tools, what's moving. No spam.