pickfit

BrowseDeveloper CLIs

agent-history

Local semantic search over AI coding agent transcripts using Ollama embeddings and SQLite.

Tested on macOS, Linux and Windows

Use it

macOS arm64

You need first

  • Runtimes: python>=3.10
  1. install via uv tool install git+https://github.com/d3layd/agent-history

    uv tool install git+https://github.com/d3layd/agent-history

Check that it works

agent-history --version 2>&1; echo ---; agent-history --help 2>&1

The output should contain “agent-history 0.16.0”.

First thing to try

agent-history inject status 2>&1; echo "---TOGGLE---"; agent-history inject toggle 2>&1; agent-history inject status 2>&1; agent-history inject toggle 2>&1

Run agent-history inject status to see whether past-session injection is turned on.

Linux x86_64

You need first

  • Runtimes: python>=3.10
  1. install via uv tool install git+https://github.com/d3layd/agent-history

    uv tool install git+https://github.com/d3layd/agent-history

Check that it works

agent-history inject --help 2>&1
agent-history inject status 2>&1; echo ---; agent-history inject off 2>&1; agent-history inject on 2>&1

The output should contain “inject: on”.

Run agent-history inject status to see whether past-session injection is turned on.

Windows x86_64

You need first

  • Runtimes: python>=3.10
  1. install agent-history via uv tool from git

    uv tool install git+https://github.com/d3layd/agent-history

Check that it works

agent-history --version

The output should contain “agent-history 0.16.0”.

First thing to try

New-Item -ItemType Directory -Force -Path docs-sample | Out-Null && Set-Content -Path docs-sample/runbook.md -Value "# Gateway Restart Runbook`nTo restart the gateway, run restart script and check auth middleware split reason.`nThe session refresh needs to run before auth check.`n" && Get-Content docs-sample/runbook.md
agent-history docs add ./docs-sample
agent-history sources

Run agent-history inject status to see whether past-session injection is turned on.

What it does

  • Search past agent transcripts by meaning
    Not tested
  • Combine meaning-based and full-text search
    Not tested
  • View the conversation around a search hit
    Not tested
  • Add relevant past sessions to your prompts
    Not shown working

How far pickfit got

macOS arm64

  1. Installs
  2. Starts
  3. 0 of 4 features work
  4. Works offline

Linux x86_64

  1. Installs
  2. Starts
  3. 0 of 4 features work
  4. Works offline

Windows x86_64

  1. Installs
  2. Starts
  3. Main function works
  4. Works offline

Good for

  • search past agent discussions by meaning
  • show conversation around a search hit
  • browse search via local web viewer
  • index documentation trees for search
  • hybrid vector and full-text search

An alternative to claude-mem.

How pickfit tested it

Each run starts from a clean machine and follows the steps above. How pickfit verifies

macOS arm64

Its main function works

  • Replayed from scratch
  • Works offline
  • Checked on 1 Nov 2026
  • Version 0.16.0

Differences the test machine hit

  • README: agent-history index works with Ollama HTTP endpoint, --mode text needs no Ollama
    In the test: index requires Ollama even for docs text-only content; needed a fake Ollama server on 11434 returning dummy 768-dim embeddings to build index before any search/show worked
  • README: prerequisites python>=3.10
    In the test: Homebrew python 3.15.2 worked (sqlite extension loadable) so no need to pin to 3.10

Linux x86_64

Its main function works

  • Replayed from scratch
  • Works offline
  • Checked on 31 Oct 2026
  • Version 0.16.0

Differences the test machine hit

  • README: agent-history index embeds everything; search --mode text needs no Ollama
    In the test: index always requires an Ollama embedding endpoint (/api/embeddings and /api/tags); in this sandbox a fake Ollama server returning deterministic 16-dim embeddings was started on 127.0.0.1:11434 to allow indexing, after which --mode text search works without vectors
  • README: uv tool install git+https://github.com/d3layd/agent-history
    In the test: prerequisites python>=3.10 and sqlite-vec loadable extension verified; doctor reported sqlite-vec: loadable on this Python 3.12

Windows x86_64

Its main function works

  • Replayed from scratch
  • Works offline
  • Checked on 31 Oct 2026
  • Version 0.16.0

Differences the test machine hit

  • README: README says `--mode text` (BM25 full-text) works without Ollama and hybrid degrades gracefully
    In the test: `agent-history index` refuses with `error: cannot reach Ollama at http://localhost:11434` when Ollama is down, so no index can be built and `agent-history search "query"` (exit 1, `error: no index ... Run: agent-history index`) cannot run without Ollama + embedding model even for --mode text
  • README: README lists `agent-history doctor` to diagnose installation without documenting exit code
    In the test: `agent-history doctor` exits 1 while still printing full diagnosis (data home, ollama up: NO, sqlite-vec: loadable, adapter/index status); exit 1 signals problems found, not a crash
  • README: (Windows steps)
    In the test: pickfit ran these steps in PowerShell 7 (pwsh 7.6.9) as a standard user; the built-in Windows PowerShell 5.1 is not verified and does not run `&&`.

How to read a tool card

Stitching
pickfit installed this tool itself. Everything under the stitches comes from that run. A tool pickfit has not run never gets stitches.
Verified on your OS
The install steps ran on the system you chose. “Ran on Linux only” means it has not been run on yours yet.
macOS, Linux and WindowsWhere it ran
Green OS marks: pickfit ran the install steps there. Grey ones: the README says it runs there, and pickfit has not tried.
How far the run got
Installs and starts, its main function works, its service answers, it produces model output. A window means a desktop app opened; a plug, an MCP client used it. The mark sits beside the run’s result, under “Why it fits and how we know” and in “How pickfit tested it”.
Visual check
pickfit looked at screenshots of the app. Green: it looks right. Amber with a number: things that look wrong; open it to see them. Grey and dashed: not checked yet.
Known vulnerabilities
Known high or critical vulnerabilities (OSV.dev) in what it installs. Green: none known. Red with a number: how many; open it for the list, and read it before you install. Grey and dashed: not checked, which does not mean safe.
About the run
The day it ran, and the version that ran.
Source of a reason
Every reason a pick fits rests on one profile field. Open the reason to see the field and the value it relies on.
Hidden gem
High quality, little known.
Quality and popularity
Two scores out of 100. Quality: how well it installed and worked when pickfit tried it, and how good its install docs and upkeep are. Popularity: from stars, downloads and Hacker News.
Features checked
Of the main features the README advertises, how many the run actually exercised (green), and how many failed (red). Grey ones were not checked, and the card says why.
Not verified
A well-known tool Fitz mentions from general knowledge. No run record, no steps, no score.