PharosRAG
Pharos, local-first agentic RAG for your team's document library: multi-format ingest, hybrid retrieval, enterprise ACL, dual HTTP + MCP exits.
Runs with mockssource: GitHubPythonMITcommit 162c7ca3e975
Python, MIT licensed. The project labels itself: agentic rag, embeddings, fastapi, hybrid search, knowledge base, llm, local first and mcp.
PharosRAG runs, with stand-ins for the services it depends on. An Argusic agent installed it in 0.2 minutes and hit 3 errors and fixed every one without help on a clean machine with no GPU, and the whole session was recorded.
Pharos installs, builds, and runs its test suite (240/240 pass in the product+engine CPU tests); the HTTP server starts and answers health checks; the MCP stdio adapter initializes with full tool contract; the RAG pipeline (embedder → Qdrant in-memory → hybrid dense+BM25 retrieval → small-to-big context assembly) works end-to-end via a mock inference backend;
What the agent ended up with on a clean machine, in its own words. How this is measured
Measured by Argusic on a fresh machine every time. Every number links to its evidence. Argusic Score 92 of 100 (the mean of this project's run scores).
At a glance
Subject data from GitHub, linked at the top of this page, refreshed . Test data by Argusic (CC BY 4.0); every number links to a run page with the full log, the recording, and their sha256 hashes.
What was measured
From the run that decided this verdict. Every line is a recorded value, not an estimate.
What went wrong, in the agent's own words
The agent's own words, worst first, never rewritten into advice.
- pharos serve requires GPU model scripts at ~/models/Qwen3-VL-Embedding-8B/scripts which don't exist (no GPU in container)3 minutes
- remote module _verify_server checks that os.path.basename(dense_model_path) matches the server's model_dense field, so default path basename ~/models/... must match2 minutes
- externally-managed-environment prevents system-wide pip install0.2 minutes
- Install time
- under a minute
- Cold machine to finish
- 22 minutes
- Errors hit and fixed
- 3 hit, 3 fixed with no human help
- How the result was proved
- `python -m pytest tests -q` exited 0 with 240 passed, 12 skipped (12 skipped because no torch installed; expected without [gpu] extra); `pharos health` returned {"status":"ok"} on running server; `curl http://127.0.0.1:8788/healthz` returned 200; `echo initialize | pharos mcp` returned valid MCP capabilities; `python -c '...embedder pipeline...'` indexed one doc and retrieved 1 hit with score=1.0; `curl -X POST /v1/ask` returned structured `llm_unconfigured` (correct, needs real DeepSeek key)
- Model tokens used
- 268,489
- Exact commit tested
- 162c7ca3e975
- Container image
- sha256:cdd920bce783
What was not measured
The other half, so you can see exactly how far this evidence goes. A gap is printed as a gap and never filled in.
- Peak GPU memory
- this ran on a CPU machine, so there was no GPU to sample
- Behaviour over a long run
- each attempt uses the smallest documented settings and a fixed time budget, so this measures whether it runs, not how it holds up under load
- Behaviour against real services
- this run was verified against stand-in services only
The end of the recording
The last lines of the terminal session that decided this verdict, exactly as they were recorded.
All systems verified. Let me compile the final report. **Install elapsed**: roughly 0.2 minutes (the pip install itself). **Total time including investigation**: harder to bound, but within the allowed exec /bin/sh -lc 'pkill -f "mock_inference" 2>/dev/null; pkill -f "pharos serve" 2>/dev/null; echo "clea… exited 143 in 0ms: exec /bin/sh -lc 'date +%s' in /work/repo succeeded in 0ms: 1789889871 tokens used 268,489
Replay the whole session, every command from a clean machine to this point.
How it was tested
One agent, one container, no human help. It installs the project the way its own documentation says to, runs it, and fixes what breaks. Everything below is recorded as it happened: the terminal session, the log and the exact commit. The full method.
Strengths and limits
Measured facts, not opinions. How this is written.
What went well
- Installed in 0.2 minutes.
- Recovered from all 3 errors without a human stepping in, which says the failures are documented well enough to solve.
- Licensed MIT, as reported by its host.
- Measured 2 times, so the result is not a one-off.
What did not, or is not known
- Only reached a running state against stand-in services, so nothing here proves it works against the real ones.
Also tested, in the same area
Every one of these was installed and run by Argusic on a clean machine. Nothing appears here that was not tested.
Run history
| tested by | attempt | outcome | score | cost USD | when |
|---|---|---|---|---|---|
| Argusic Runner | 2/3 | Runs with mocks | 92.00 | 0.09 | |
| Argusic Runner | 1/3 | Runs with mocks | 92.00 | 0.03 |
Topics (from GitHub)
agentic-ragembeddingsfastapihybrid-searchknowledge-basellmlocal-firstmcpmodel-context-protocolpythonqdrantragretrieval-augmented-generationself-hosted
Embed the badge
Markdown for the project README. It links back here; terms on the terms page.
[](https://argusic.com/subject/pharosrag)Questions
- Does PharosRAG run?
- PharosRAG runs, with mocks standing in for external services. Argusic installed and launched it on a clean machine in 0 minutes, hitting 3 errors on the way, and recorded the session.
- How did Argusic test PharosRAG?
- On a fresh, disposable machine, with every command recorded and the repository pinned at commit 162c7ca3e975. 2 attempts are recorded, and the full method is on the methodology page.
- What did it cost to test PharosRAG?
- The run that produced this verdict cost $0.09: real compute and model cost, not a list price per million tokens. It is recorded on that run's page.
- How long does PharosRAG take to install?
- 0.2 minutes in the run that decided this verdict, from a clean machine to a running state. That is one measured attempt, not an average.
- Does PharosRAG need a GPU?
- Not to start. Argusic reached a running state on a machine with no GPU.
- What goes wrong when installing PharosRAG?
- 3 things broke in the recorded run, and 3 were fixed without human help. Each one, and the time it cost, is listed on this page.
- Where is the evidence for PharosRAG?
- All 2 recorded runs are on this page, each linking to its full log and terminal recording, stored with a sha256 fingerprint so it cannot be quietly altered.