Library · Team · Ollama · deep-research
Deep Research — Ollama
Turn a fuzzy question into a sourced research brief with open questions and next checks — not a final verdict. (Ollama Modelfile + chat prompt for offline runs.)
Id: ollama/deep-research · Slug: deep-research · Status: Verified · 2026-09-08 · Real Ollama run PASS on 2026-09-08T18:04:14.904670+09:00. ollama version is 0.33.3; model botshelf-deep-research:latest (llama3.2:1b). Evidence: qc/acq-sales/VERIFY-deep-research-ollama-2026-09-08.md. Limitations: 1.2B model; placeholder me
What is this
Prompt for Ollama. Turn a fuzzy question into a sourced research brief with open questions and next checks — not a final verdict. (Ollama Modelfile + chat prompt for offline runs.)
How this runtime fits
This Ollama recipe is a local AI agent / local AI workflow for the parent AI Team. It is a single-prompt (or system-prompt) recipe — do not force multi-agent system language here.
Parent AI Team: deep-research · What is an AI Team?
What it does
Turn a fuzzy question into a sourced research brief with open questions and next checks — not a final verdict. (Ollama Modelfile + chat prompt for offline runs.)
Who for
Builders working on Research jobs who can run Ollama themselves.
Need to run
- You will run this locally or on infra you control.
- No production credentials in prompts or committed files.
- Ollama installed
- A local instruct model pulled
How to use
- Install Ollama and pull a base model you already trust (example: llama3.2).
- Save the Modelfile above in an empty folder.
- Run: ollama create botshelf-deep-research -f Modelfile
- Run: ollama run botshelf-deep-research
- Paste your input (see Example in). Expect the structured output; then stop.
- Note: llama3.2:1b showed invent/numeric failures in Proof Engine — prefer full llama3.2 (3B+) for coding-review/data-analysis.
Limitations
- Not a substitute for professional legal, medical, or investment advice.
- Outputs can be wrong; human review required before publish or spend.
- Stage-1 Library items are free recipes — marketplace ready-to-use teams remain separate.
- Local models may omit sections; re-prompt once if needed.
Test status
Status: Verified · 2026-09-08 · Real Ollama run PASS on 2026-09-08T18:04:14.904670+09:00. ollama version is 0.33.3; model botshelf-deep-research:latest (llama3.2:1b). Evidence: qc/acq-sales/VERIFY-deep-research-ollama-2026-09-08.md. Limitations: 1.2B model; placeholder me
Files
Modelfile— Create a local model with the system promptprompt.md— User-facing run instructions
Full prompt / config / code
Free. No sign-up. The full text is right here. Copy it, or take the file with you.
無料・登録不要。全文はこの下にあります。コピーするか、ファイルで持ち帰れます。
Download .md.md で持ち帰る How to run it動かし方 Fork itフォークする Results & questions結果・質問を書く
# Modelfile — Deep Research (Ollama) FROM llama3.2 SYSTEM """ You are a careful research desk. Goal: produce a brief, not a verdict. RULES: - Use only facts present in the user paste or clearly marked as [UNVERIFIED]. - Separate: Known / Inferred / Unknown / Next checks. - Label every finding Known vs Inferred. Never state Inferred as fact. - Do not invent causes (trust, reputation, criteria changes, market effects) not in the paste. - Do not invent metrics, percentages, or dates. - If recommending a priority, phrase as hypothesis under Open questions / Next checks — not as decided scope. - Prefer quoting pasted labels/counts exactly. - Cite sources as the user labeled them (URL, doc name, date). Do not invent citations. - No spending, no shell commands, no credential requests. - Stop after the brief. Ask one clarifying question only if blocking. OUTPUT: 1) One-sentence scope 2) Findings (bullets) 3) Open questions 4) Next checks (max 5) 5) Confidence: low|medium|high + why """ PARAMETER temperature 0.3 PARAMETER num_ctx 8192 # Build & run (safe, local): # ollama create botshelf-deep-research -f Modelfile # ollama run botshelf-deep-research # # Model note: llama3.2:1b showed invent/numeric failures in Proof Engine; # prefer full llama3.2 (3B+) for coding-review/data-analysis. Keep FROM llama3.2 # as-is (matches Library). # # Safety: no network tools required. Do not pass API keys into the Modelfile.
License as listed in the registry: free-use-at-own-risk. Use is at your own risk. BSV does not run it for you. What is open here
ライセンス(レジストリの記載どおり):free-use-at-own-risk。 利用は自己責任です。BSVが代わりに実行することはありません。何が開いているか
This job is also written for 6 other tools. See all 7 versions side by side.
この仕事は、ほかに 6 のツール向けの版があります。7 版を並べて見る。
Example in
Question: Why did our signup conversion drop last week? Sources: - [A] analytics export 2026-09-01..09-07 (CSV summary pasted) - [B] changelog: checkout button copy tweak on 09-03 Ask for a brief with next checks.
Example out
Scope: Explain likely drivers of signup conversion drop using only pasted analytics + changelog. Findings: - Drop concentrated on mobile checkout step (from A) - Timing aligns with 09-03 copy tweak (B) Open questions: - Was A/B still running? Next checks: 1) Diff mobile CTA copy pre/post 2) Error rate on payment iframe Confidence: medium — single-week window.
Model / runtime notes
FROM llama3.2 (Library default). llama3.2:1b showed invent/numeric failures in Proof Engine; prefer full llama3.2 (3B+) for coding-review/data-analysis. Works best with 8k+ context instruct models. Smaller models may skip sections — ask it to continue once.
Canonical Team
Parent: deep-research
· implementation_id: ollama/deep-research
GitHub source
teams/deep-research/ollama (path in the BSV team registry repository) · commit 84776c746c5144e03556db48141ab718c3bd2651
· license: free-use-at-own-risk
Runtime verification: ollama local box; ollama version is 0.33.3; OLLAMA_MODELS=/workspace/tools/ollama-models
Related Library items
Related marketplace Team pages
Soft thematic links only — not identity merges. Marketplace Teams ≠ Library AI Team records.
- Research Desk (marketplace Team page — separate entity)
- Source Desk (marketplace Team page — separate entity)