← Default view
ATLAS / OBSIDIAN-2026 / AI-PLUGINS / DASHBOARD
SURVEY · 20 SRC 5 MIN READ APR 2026
CONSOLE / OBSIDIAN AI PLUGINS / IN-VAULT RETRIEVAL · SUMMARIZATION

Four plugins worth installing.
One you should never run together.

The 2026 ecosystem split into clear lanes — retrieval, chat-with-vault, editor-grade writing, cross-device second brain. Most users only need one from each of the first two columns. Smart Connections is the local-first retrieval default; Copilot is the polished chat; Smart Composer is Cursor-in-Obsidian; Khoj wins when "vault" extends past Obsidian. Everything else is niche or duplicative.

PLUGINS SURVEYED
12
4 main · 3 summarization · 5 niche
COMBINED ⭐
50.3k
khoj 34k · copilot 6.8k · sc 4.9k · sm-comp 2.2k
FREE-TIER COVERAGE
100%
all four ship a working free tier
FIRST-INDEX TIME
20–30min
5,000-note vault, local embeds[15]
DOUBLE-INDEX RISK
SC + Copilot QA + SC RAG = 3 indexes[13]

Plugin Tiles · Top Picks SORTED BY ROLE

Smart Connections preview RETRIEVAL ⭐ 4.9k
Smart Connections
brianpetro · obsidian-smart-connections

The only plugin here that ships local on-device embeddings with zero config and keeps working offline once indexed[2].

EMBEDSlocal, on-device
RAGSmart Chat (free)
EDITS— retrieval only
AGENTvia MCP[19]
PRICE · FREE core · Pro via supporter sub[3]
→ github.com/brianpetro/…
Copilot for Obsidian preview VAULT CHAT ⭐ 6.8k
Copilot for Obsidian
logancyang · obsidian-copilot

Most polished local-LLM chat experience — native Ollama, "Relevant Notes" auto-context, broadest multimodal handling[14].

EMBEDSopt-in (Ollama)
RAGVault QA mode
EDITScommand palette
AGENTPlus tier only
PRICE · BYOK free · Plus $14.99/mo · self-host $349.99 once[5]
→ github.com/logancyang/…
Smart Composer preview EDITOR + CHAT ⭐ 2.2k
Smart Composer
glowingjade · obsidian-smart-composer

Cursor-in-Obsidian. @file mentions feed exact files into context, @Vault triggers RAG, Apply-Edit accepts rewrites with one click[8].

EMBEDSvia Ollama / LM Studio
RAG@Vault · ⌘⇧⏎
EDITS✓ one-click Apply
AGENTMCP tools
PRICE · FREE with own keys/subs · Gemini API works as free tier[7]
→ github.com/glowingjade/…
Khoj preview SELF-HOST ⭐ 34k
Khoj
khoj-ai · khoj

A self-hostable second brain reachable from Obsidian, Emacs, browser, desktop, phone, or WhatsApp — the most actively maintained of the bunch[9].

EMBEDSself-hosted server
RAGchat over docs + web
EDITS— chat-only
AGENTcustom agents
PRICE · FREE open-source self-host · managed cloud tiers[9]
→ github.com/khoj-ai/…

Feature Matrix Console ● YES  ◯ OPTIONAL  ✗ NO

Feature Smart Connections Copilot Smart Composer Khoj
Local embeddings out of box zero-config offline on-device model[2] Ollama + nomic-embed-text[6] via Ollama / LM Studio[7] self-hosted server[9]
LLM provider span cloud + local local + 100+ APIs[1] OpenRouter, OpenAI, Anthropic, Gemini, Cohere, Ollama, LM Studio[4] OpenAI, Anthropic, Gemini, Groq, DeepSeek, OpenRouter, Azure, Ollama, LM Studio[7] gpt, claude, gemini, llama, qwen, mistral[9]
RAG / vault chat conversational Q&A Smart Chat (free in core) Vault QA mode @Vault · ⌘⇧⏎[8] chat over docs + web[10]
Apply-Edit / inline rewrites writer surface retrieval only command palette quick actions one-click Apply chat-only
Multimodal context YouTube · PDF · images YouTube, web, PDF, EPUB, images websites, images, YouTube transcripts PDF, image, Word, Notion, org-mode
Agent / tool-use autonomous calls external via MCP[19] Plus tier: Agent mode MCP support custom agents, automations
Privacy posture cloud vs local default local-first, offline after index local with Ollama; cloud by default local with Ollama; cloud optional self-host or Khoj cloud
Pricing floor → ceiling Free core · Pro via sub[3] BYOK free · Plus $14.99/mo · $139.99/yr[5] Free with own keys Free OSS · managed cloud tiers

Ecosystem Lanes SIX LANES · TOP PICK PER LANE[12]

L01RETRIEVAL / DISCOVERY

Embeds every note locally, surfaces related ones in a sidebar

A Lookup view + related-notes pane while you write. Local-first, offline after first build.

PICK · Smart Connections[2]
L02CHAT WITH VAULT (RAG)

Conversational Q&A grounded in your notes via embeddings + LLM

Three viable choices. Pick exactly one — never run two embedding indexes over the same vault.

PICK · Copilot[4], Smart Composer[7], Khoj[10]
L03EDITOR-GRADE WRITING ASSIST

@file mentions, Apply-Edit, inline rewrites

Cursor-style flow inside Obsidian. The chat-and-write boundary blurs.

PICK · Smart Composer[8]
L04TEMPLATE-DRIVEN GENERATION

Reusable prompts (summary, outline, expand) tied to frontmatter

Ships GPT-5 + thinking-model support; community template gallery.

PICK · Text Generator[11]
L05DEDICATED SUMMARIZATION

One-shot summarize commands with chunking

Mostly redundant in 2026 — Copilot, Smart Composer, and Text Generator already cover this. Two niche cases below.

NICHE · obsidian-ai-summary[16], obsidian-ai-summarize[17]
L06AGENTIC (READ + EDIT + CREATE)

Autonomous tool calls that modify the vault

Interesting but immature. Reliable 2026 pattern: wire Smart Connections retrieval into an external agent via MCP[19].

EARLY · Obsidian Chat, Vault AI Chat[20]

Operational Warnings 4 GOTCHAS · READ BEFORE INSTALLING

⊠ HARD RULE

Don't double-index

Running Smart Connections + Copilot Vault QA + Smart Composer RAG simultaneously means three separate embedding indexes over the same vault[13]. Pick one chat plugin; let Smart Connections handle the discovery sidebar.

⏱ FIRST-RUN COST

20–30 min on a 5,000-note vault

First-build embedding indexing locally takes 20–30 minutes on a 5k-note vault. Index lives in .obsidian/ and updates incrementally afterward[15]. Plan for the first run, not the steady state.

▣ KNOWN-GOOD STACK

Local-only: Ollama + nomic-embed-text + Qwen 2.5

2026's known-good local stack: Ollama + nomic-embed-text for embeddings + Qwen 2.5 (or Llama 3) for chat, wired into Copilot or Smart Composer[6][14].

⚠ IMMATURE LANE

Agentic plugins are early

Plugins like Obsidian Chat and Vault AI Chat can read+edit+create files autonomously[20] — interesting but immature. The more reliable 2026 pattern: wire Smart Connections retrieval into Claude Code or another external agent via MCP[19].

Summarization Niche SKIP UNLESS YOU MATCH ONE OF THESE

obsidian-ai-summary

⭐ 39

Walks links from the current note and summarizes each linked note in a dialog[16].

USE WHEN · weekly/monthly review note walking all linked notes
→ irbull/obsidian-ai-summary

obsidian-ai-summarize

⭐ 18

One-shot summarize command with chunking + placement profiles[17].

USE WHEN · simpler UX than Text Generator, smaller scope
→ ravenwits/obsidian-ai-summarize

Text Generator

⭐ 1.9k

Reusable summary templates with frontmatter overrides; community gallery, GPT-5 + thinking-model support[11].

USE WHEN · template-driven generation tied to frontmatter
→ nhaouari/obsidian-textgenerator-plugin

Install Order THE TWO-PLUGIN BASELINE + OPTIONAL ADDS

01
Install Smart Connections

Local-first sidebar. Zero config. Index first. Offline after.

02
Pick ONE chat plugin

Copilot for polish · Smart Composer for editing · Khoj if cross-device.

03
Stand up Ollama

Pull nomic-embed-text + Qwen 2.5 / Llama 3. Point chat at localhost:11434.

04
Resist double-indexing

Disable any second RAG index. Smart Connections owns discovery.

05
(Optional) Wire MCP

Expose retrieval to Claude Code or external agents. The 2026 advanced pattern.

Sources Console · 20 cited

[01]
brianpetro/obsidian-smart-connections
Smart Connections — chat with notes via AI embeddings; local + 100+ API providers.
OFFICIAL · ⭐ 4.9k
[02]
smartconnections.app
Indexing and retrieval run locally by default; offline after first index.
VENDOR-BLOG
[03]
smartconnections.app/pro-plugins
Core remains free; Pro tier with 14-day trial yearly or monthly plan.
VENDOR-BLOG
[04]
logancyang/obsidian-copilot
Vault-aware chat with Vault QA / RAG, @ mentions, Ollama support, Plus-tier Agent mode.
OFFICIAL · ⭐ 6.8k
[05]
obsidiancopilot.com/en/pricing
Plus $14.99/mo or $139.99/yr; one-time $349.99 Self-Host Supporter.
OFFICIAL
[06]
obsidian-copilot/local_copilot.md
Run fully local with Ollama at localhost:11434 + nomic-embed-text.
OFFICIAL
[07]
glowingjade/obsidian-smart-composer
Cursor-style chat with @file/@vault, Apply-Edit, MCP support across many providers.
OFFICIAL · ⭐ 2.2k
[08]
smart-composer/wiki/2.3-Vault-Search-(RAG)
Vault Search RAG triggered via Cmd+Shift+Enter or @Vault.
OFFICIAL
[09]
khoj-ai/khoj
Self-hostable AI second brain accessible from Obsidian, browser, phone, WhatsApp.
OFFICIAL · ⭐ 34k
[10]
docs.khoj.dev/clients/obsidian
Khoj Obsidian plugin: chat, transformer-based incremental search, similar-notes discovery.
OFFICIAL
[11]
nhaouari/obsidian-textgenerator-plugin
Template-driven generation across OpenAI, Anthropic, Google, HuggingFace, local.
OFFICIAL · ⭐ 1.9k
[12]
systemsculpt.com — Best Obsidian AI 2026
April 2026 review notes the ecosystem now splits into retrieval/workflow/agent surfaces.
VENDOR-BLOG
[13]
forum.obsidian.md — alternatives
Copilot's Vault QA indexes notes for semantic search similar to Smart Connections.
FORUM
[14]
xda-developers.com — local LLM with Obsidian
XDA recommends Copilot as the most polished local-LLM option; Relevant Notes auto-context.
NEWS
[15]
insiderllm.com — Obsidian local LLM guide
First-build embedding indexing on 5k-note vault: 20–30 min locally.
VENDOR-BLOG
[16]
irbull/obsidian-ai-summary
Walks links from current note, summarizes each linked note in a dialog.
OFFICIAL · ⭐ 39
[17]
ravenwits/obsidian-ai-summarize
Small dedicated summarization with chunking and placement profiles.
OFFICIAL · ⭐ 18
[18]
danielrosehill/awesome-obsidian-ai-tools
Community-maintained list of retrieval, summarization, and agent plugins.
OFFICIAL · ⭐ 115
[19]
3sztof.github.io — SC + MCP
Wiring Smart Connections retrieval into a full agent loop via MCP.
FORUM
[20]
forum.obsidian.md — Obsidian Chat agentic
Newer agentic plugins read notes, edit, and create files autonomously.
FORUM

Sibling Angles · Obsidian 2026 Expedition

VIEW · DASHBOARD CANONICAL · ↩ DEFAULT VIEW TOPIC · OBSIDIAN AI PLUGINS · APR 2026 20 SOURCES · 5 MIN