Researchers & scientists — real builds and copy-ready starters, grouped by what you’re trying to make. Filter by type (web app vs. agent) or tool, or search.
v0 Scientist AI-native SMEA chart-driven UI where filters and a dropdown re-render the visualisation against a dataset.
A hands-on way to let people poke at your data instead of reading a static chart.
Create an interactive data explorer with a metric dropdown, a date-range filter, and a category selector that together drive a responsive line chart and a sortable data table beside it.
Hermes Small biz ScientistThirteen finance skills covering Hong Kong, US and A-share markets. Needs the finance-skills pack installed first.
Live market data and analysis on tap for a solo trader or analyst.
Pull today’s closing prices for [my HK, US or A-share tickers] and send me a daily summary on Telegram.
A workflow splits a labelled dataset, trains a Random Forest classifier using KNIME's learner nodes, evaluates it with a scorer, and deploys it as a local REST endpoint for batch prediction.
A working predictive model is built and validated without writing any code, and retraining is a one-click re-run.
CSV Reader → Partitioning node (80/20) → Random Forest Learner → Random Forest Predictor → Scorer → ROC Curve node → KNIME Server REST endpoint (deploy).
OpenClaw AI-native SME Founder ScientistOpenClaw polls an internal API or public endpoint on a schedule, evaluates a condition you define (error rate, response time, a numeric threshold), and fires an alert to Telegram or Slack the moment the condition is breached — along with the raw value and a short triage note.
You are notified the instant a KPI or system metric crosses a threshold, without running a monitoring stack or paying for an alerting service.
Message your bot: "Every 15 minutes, hit [endpoint]. If the response time is over 2s or the status is not 200, ping me on Telegram with the value and a one-line triage note."
Gemini ScientistPaste a table into Gemini, ask it to describe the trends, and have it generate a chart and a figure illustration for a slide.
A quick description plus a draft visual, using Gemini's multimodal strengths.
Paste a CSV snippet → "Describe the trend, then generate a labeled bar chart and a simple schematic figure."
Base44 Creator ScientistA portal where students enroll in a course, work through lessons, submit assignments, and see their progress and grades.
Run a cohort course with submissions and grading without paying for a heavy LMS.
Build a course portal where students log in, work through lessons, and submit assignments. Let the instructor grade submissions and leave feedback. Show each student their own progress bar and grades, and give the instructor a class-wide dashboard.
Lovable AI-native SME ScientistAn internal booking system for shared rooms or instruments with conflict-free scheduling.
End the double-booking arguments over the shared microscope or meeting room.
Build a booking system for shared resources like meeting rooms or lab instruments. Show each resource’s calendar, let staff reserve a time slot, prevent overlapping bookings, and give admins a view of all reservations with the ability to cancel.
v0 Scientist AI-native SMEAn internal scheduler for shared instruments showing availability, a booking form and a calendar view.
Stops the spreadsheet chaos around shared lab gear.
Create a lab equipment booking tool with a list of instruments, a weekly calendar showing existing bookings, a slot-reservation form with user and purpose fields, and a “my bookings” panel.
Codex reads an existing analysis library and generates a Click-based CLI wrapper with subcommands for each public function, help text derived from docstrings, and integration tests — all added to the repo without touching the library itself.
Lab members run analyses from the terminal without writing any Python; the CLI ships as part of the next package release.
Run `codex`: "Add a Click CLI in cli/main.py that exposes all public functions in src/analysis/pipeline.py as subcommands. Derive help text from existing docstrings. Add tests in tests/cli/."
Ollama AI-native SME ScientistA short Python script loops over HR or patient records, sends each row to a locally running Llama 3 model via the Ollama REST API, and asks it to replace names, addresses, and ID numbers with placeholders — the cleaned file is written locally before anything goes to an external service.
Sensitive records are de-identified on the machine, satisfying data-governance requirements without involving a third-party API.
Run `ollama pull llama3`, then POST each row to `localhost:11434/api/generate` with prompt: "Replace all PII in the following text with [REDACTED_<TYPE>]. Return only the cleaned text."
Claude Scientist AI-native SMEAsk Claude to help you create a structured markdown log — one entry per session — recording which task was delegated, what the model produced, and what human verification was done. Commit the log alongside your project files.
If a co-author, editor, or auditor asks "where exactly did AI help?", you have a traceable record rather than a vague recollection.
Prompt: "Create a template for an AI-use log with columns: Date, Task, Model, Output summary, Verification step taken."
v0 ScientistAn academic group site with sections for research themes, publications, team members and news.
A credible lab presence without fighting an old university CMS.
Create a university research lab website with a hero stating our research focus, a grid of research themes, a publications list with filter by year, a team page with photo cards, and a news feed. Clean and academic.
Hermes Creator ScientistA built-in skill that pulls a YouTube transcript and turns it into a summary or content outline.
Use the youtube-content skill to pull the transcript from this video [paste URL], summarise the key points, and draft three tweet ideas based on it.
LM Studio ScientistA researcher uses LM Studio's chat interface on a field laptop without internet to draft specific grant sections — significance, innovation, approach — by feeding the model their notes and asking for structured prose that matches funder guidelines.
Grant drafts progress during travel or remote fieldwork with no connectivity dependency and no risk of unpublished ideas reaching a cloud provider.
In LM Studio, load a strong instruction model (e.g. Qwen 2.5 14B), paste your bullet-point notes, and prompt: "Write the Significance section of an NIH R01 application based on these notes. Use formal scientific register."
Base44 ScientistA laboratory tracker for samples and reagents — each item has a location, owner, expiry date, and full status history.
No more lost samples or expired reagents discovered mid-experiment; the whole lab queries one searchable database.
Build a lab inventory app to track samples and reagents. Each item should have a name, freezer location, owner, quantity, and expiry date, with a log of who took what and when. Give each lab member a login and alert me when a reagent is low or expiring.
Base44 ScientistA portal where study participants register, book session slots, complete consent forms, and view their schedule, while researchers manage everyone from an admin view.
Cuts the email back-and-forth of recruiting and scheduling a study cohort into one self-serve system.
Build a research participant portal where people sign up, fill in an eligibility questionnaire, sign a consent form, and book an available study slot. Give researchers an admin dashboard to approve participants and see the full schedule.
Base44 ScientistA digital lab notebook where researchers log experiments with protocol, conditions, results, and attached data files.
A searchable, shareable record of every experiment that survives staff turnover.
Build an electronic lab notebook where researchers log experiments with a title, protocol steps, conditions, observations, and attached result files. Make entries searchable, give each scientist a login, and let them link related experiments together.
Base44 ScientistA shared library where a research group logs papers they’ve read, tags topics, and writes a short takeaway for each.
The whole lab’s reading becomes collective knowledge instead of scattered PDFs.
Build a reading tracker for my research group. Add papers with title, authors, link, topic tags, and a one-paragraph takeaway. Let members log in, mark papers as read, and filter by topic. Show who has read the most this month.
Base44 ScientistA portal where authors submit abstracts, reviewers score them, and organizers see the ranked list and accept or reject.
Runs a small conference’s review process without expensive dedicated software.
Build an abstract submission portal. Authors log in and submit an abstract with title, summary, and topic. Assign reviewers who score each submission 1–5 with comments. Give organizers a dashboard ranking abstracts by average score with an accept/reject action.
Base44 ScientistA tracker for a longitudinal study — each participant has scheduled visits, data-collection status, and follow-up reminders.
Keeps a multi-visit study on schedule and flags participants who are overdue.
Build a study cohort tracker. Each participant has a schedule of visits with dates and a status for whether data was collected. Remind me of upcoming and overdue visits, restrict access to authorized researchers, and show enrollment and retention on a dashboard.
Base44 Scientist AI-native SMEA mobile-friendly app where field workers record observations with photos, GPS notes, and timestamps that sync to a central database.
Replaces paper field sheets with structured, instantly-shared data the whole team can analyze.
Build a field data collection app where workers log in on their phone and record observations with a category, notes, photo, and location. Sync everything to a central database and give the lead a dashboard and map-style table of all entries.
Lovable ScientistA portal where study participants log in, consent, complete scheduled surveys, and view their progress.
Run a longitudinal study without paying for enterprise survey software.
Build a research participant portal where people sign up, e-sign a consent form, and complete a series of timed surveys. Save every response to the database with a timestamp, show participants which sessions they’ve finished, and give me an admin view to export all responses as CSV.
Lovable Scientist AI-native SMEAn internal tool to track reagents, equipment, and stock levels with low-stock alerts and a checkout log.
Replace the shared spreadsheet that nobody keeps up to date.
Build a lab inventory tracker where team members log in and record reagents and equipment with quantity, location, and expiry date. Let people check items in and out, flag anything below a minimum threshold, and show a dashboard of low-stock and soon-to-expire items.
Lovable ScientistA form-driven app for collecting field observations with photos, location, and offline-friendly entry.
Standardize field data capture across a whole research team.
Build a field data collection app where researchers log in and submit observations with a category, notes, GPS coordinates, and a photo upload. Store each entry with the collector’s name and timestamp, show all entries on a map, and let me export the dataset.
Lovable ScientistA digital lab notebook where researchers record experiments with protocols, results, and attachments.
Keep a searchable, timestamped record the whole team can trust.
Build a shared experiment log where researchers log in and create entries with a title, protocol, parameters, results, and file attachments. Timestamp each entry, make them searchable, let people link related experiments, and keep an edit history.
v0 ScientistA participant-facing page explaining a study, eligibility criteria and a screening signup form.
A trustworthy front door for recruiting study participants.
Create a research study recruitment landing page with a clear plain-language summary, an eligibility checklist, what participation involves, compensation info, and a short screening signup form. Calm, trustworthy tone.
v0 ScientistA shareable page for a single paper or poster: abstract, figures, authors, and links to data and code.
A modern alternative to a static PDF when sharing your work.
Create a microsite for a single research paper with the title and authors, a structured abstract, a figures gallery with captions, key findings as highlight cards, and buttons to the PDF, dataset and code repository.
Hermes ScientistA daily agent on a lab workstation pointed at a local model summarises each morning’s new PubMed alerts and learns which topics you actually care about.
Zero-cost and fully local — abstracts and queries never leave the lab, and it gets more relevant the longer you use it.
Every weekday at 8am, search PubMed for new papers on my saved topics, summarise the 5 most relevant, message me the digest, and weight future digests toward the ones I open.
Hermes ScientistHermes-Lab: give it a search space and an evaluation method and it schedules runs, tracks every result, and suggests what to try next. Needs the Hermes-Lab skill installed first.
The bookkeeping of an ML sweep runs itself; you just decide the next direction.
Here is my search space and evaluation function — schedule runs, track every result in memory, and message me on Telegram with your next suggested experiment after each batch.
Hermes ScientistChEMBL, AlphaFold, OpenFDA and QSAR workflows wired into Hermes skills to run early drug-discovery queries from a laptop. Needs the drug-discovery toolkit skill installed first.
Heavyweight computational-chemistry workflows become a chat you can run anywhere.
Using ChEMBL and OpenFDA, find approved compounds similar to [your target molecule], run a QSAR filter, and save the shortlist as a skill I can re-run weekly.
Hermes ScientistWalk Hermes through cleaning one messy instrument-export CSV; it writes a reusable skill the first time and applies it automatically next week.
Teach a fiddly procedure once; the agent owns it from then on and gets faster every run.
Help me clean this instrument CSV step by step. When we're done, save what we did as a reusable skill so you can clean the next export the same way without me re-explaining.
Hermes ScientistA retrieval setup (PageIndex) where the agent understands a document's structure and reasons over it with tools instead of chopping it into vector chunks. Needs the PageIndex skill installed first.
Answers that respect how a paper is actually organised, with fewer hallucinated links.
Index these papers from my folder and answer [your research question] by reasoning over their structure with tools — no vector chunks.
Hermes ScientistA data scientist ported an entire Python weather stack (MetPy, Herbie, cfgrib, WRF-Python) to Rust and built two Hermes plugins around it. Needs the weather-stack plugin installed first.
A slow, fragile scientific pipeline becomes a fast, agent-drivable tool.
Run my MetPy/Herbie pipeline on today’s GFS data, save the output plots to my results folder, and schedule this every morning at 6am.
Hermes ScientistA research stack of seven MCP tools plus a self-hosted SearXNG search to replace a paid research API.
Deep web research on your own infrastructure, squeezing the free tiers instead of paying per query.
Build me a private research engine using SearXNG plus web, PDF, and citation tools — save it as a skill called deep-research so I can call it with any question and get a cited summary.
Hermes Scientist Creator AI-native SMEWatches your field, picks the real signals, writes a brief, suggests angles, and delivers via Discord, Slack, Notion, email, Obsidian or markdown.
One self-improving morning brief instead of ten tabs you never get to.
Each morning, scan my field's top sources, pick the 5 developments that matter, write a short sourced brief, and post it to my Slack. Learn which items I react to.
Hermes ScientistA community skill that turns Hermes into an autonomous research agent for surveying and summarising a topic. Needs the literature-research-agent skill installed first.
A first-pass literature scan that runs while you do something else.
Survey and summarise the last two years of papers on [your topic], save the summary to memory, and message me the key findings on Telegram.
Hermes Scientist CreatorReads Hacker News daily and emails a curated summary of what's worth your attention.
The signal from a noisy feed, on a schedule, without the doomscroll.
Every morning at 7am, read Hacker News, filter for science and engineering stories relevant to [your field], and email me a curated brief with your picks and one-line summaries.
Hermes ScientistA starter where Hermes records every analysis step you run into a dated markdown notebook and can replay or summarise a day's work on request.
Reproducibility for free — the record writes itself as you work.
From now on, after each analysis step, append what we did and why to today’s lab notebook. At 6pm, post me a summary of everything we ran today.
Hermes ScientistA built-in skill to search arXiv — point a cron at it to surface new papers on your topics.
Every Monday at 8am, use the arxiv skill to search for new papers on [my topic], summarise the 5 most relevant, and message me the list.
Hermes ScientistA built-in research-paper-writing skill that structures an ML paper for a conference from your results.
Use the research-paper-writing skill to draft a conference-style ML paper from my results, structuring it with abstract, related work, methodology and evaluation sections.
Hermes ScientistA built-in skill that logs runs and builds dashboards in Weights & Biases.
Use the weights-and-biases skill to log my current experiment’s hyperparameters and metrics, then build a dashboard comparing runs by validation loss.
Hermes ScientistBuilt-in skills to run local inference (vLLM for throughput, llama.cpp for GGUF) so the whole group queries one private model.
Use the vllm skill to spin up a local inference server for [my chosen model], then share the endpoint URL with my team so everyone queries one private instance.
Hermes ScientistA built-in skill that runs Python step-by-step against a live Jupyter kernel, keeping state between cells.
Use the jupyter-live-kernel skill to load my dataset, run an exploratory analysis step-by-step in a live kernel, and keep the variable state ready for follow-up queries.
Hermes ScientistA built-in Segment Anything skill for segmenting microscopy or imaging data without training.
Use the segment-anything-model skill to segment the regions of interest in my microscopy images and export the masks as labeled PNG files.
Hermes ScientistA built-in skill to pull, push and manage Hugging Face models and datasets from the agent.
Use the huggingface-hub skill to pull the latest checkpoint for my fine-tuned model, evaluate it on my test split, and push the updated weights back to my repo.
Hermes ScientistA built-in skill to run LLM evaluation harnesses and report scores.
Use the evaluating-llms-harness skill to benchmark my local model on MMLU and HellaSwag, then save the score report to a timestamped file.
Hermes Scientist CreatorA built-in skill that monitors blogs and RSS/Atom feeds and flags what is new and relevant.
Use the blogwatcher skill to monitor my five favourite ML blogs and Nature RSS feeds, then message me on Telegram whenever a post matches my research keywords.
Hermes ScientistA community plugin exposing National Weather Service models and NEXRAD radar imagery to the agent. Needs the hermes-weather-plugin skill installed first.
Pull the latest NWS forecast and NEXRAD radar imagery for [my study region] and summarise current conditions.
Hermes ScientistHermes drove TurboQuant to optimise an MLX model (Qwen3.5-9B) for local inference on a Mac.
Drive TurboQuant to quantize [my target model] for Apple Silicon, benchmark inference speed before and after, and save the optimised MLX weights locally.
Hermes ScientistUsing Hermes to train and evaluate tool-calling models against a 100-task benchmark.
Set up my tool-calling model training loop, run it against the 100-task benchmark suite, and message me on Telegram with the final pass-rate scores.
Hermes ScientistA single question fanned out across Reddit, X, YouTube, Hacker News and Polymarket, then synthesised.
Fan out my research question '[my topic]' across Reddit, X, YouTube, Hacker News and Polymarket, then synthesise the key signals into a one-page brief.
Hermes ScientistA community project running a Mars rover simulation through ROS2 and Gazebo, driven by Hermes. Needs the Hermes-mars-rover community skill installed first.
Run a navigation scenario in the ROS2/Gazebo Mars-rover simulator and log the rover’s path to a CSV.
Hermes ScientistA community project fine-tuning vision-language-action models so the agent improves a robot over time. Needs the hermes-embodied community skill installed first.
Run one training iteration on my robot demonstration data using the vision-language-action fine-tuning skill and report the improvement metric.
n8n ScientistEvery Monday morning the workflow queries PubMed for new papers matching your MeSH terms, sends each abstract through an AI summarize node, and posts a ranked digest to your lab Slack channel.
The lab gets a curated 10-paper reading list in Slack before the weekly meeting — no manual search needed.
Schedule trigger (Monday 08:00) → HTTP Request node (PubMed E-utilities API, search + fetch abstracts) → AI node (summarize each abstract to 2 sentences) → Slack node (post formatted digest).
A KNIME workflow reads raw plate-reader exports, removes outlier wells, normalises to controls, calculates mean and CV per condition, and writes a tidy results table plus a bar-chart figure.
A 45-minute manual Excel clean-up is replaced by a reproducible, one-click pipeline any lab member can run.
CSV Reader → Missing Value node → Row Filter (remove flagged wells) → Math Formula (normalise to DMSO control) → GroupBy (mean, SD, CV per condition) → Box Plot node → Excel Writer.
Reads a multi-sequence FASTA, loops over each entry, queries the UniProt REST API for annotation, and assembles a flat table of sequences with GO terms, organism, and reviewed status.
Hundreds of sequences are annotated overnight without writing a single line of Python.
File Reader (FASTA) → Chunk Loop Start → REST Client node (UniProt API, GET by accession) → JSON Path → Loop End → Column Appender → CSV Writer.
Merges visit records from multiple CSV exports, checks for missing values and out-of-range measurements, flags anomalies, and produces a formatted QC report with per-site summaries.
Every data transfer is quality-checked consistently and a report is ready before the weekly data review call.
Multiple CSV Readers → Joiner (merge on participant ID) → Missing Value node → Rule Engine (flag out-of-range) → GroupBy (QC stats per site) → Report Template node.
Dify ScientistResearchers upload a batch of PDFs from a systematic review, then ask cross-cutting questions ("which studies used a randomised design?") and get synthesised answers with paragraph-level citations.
Early-stage literature synthesis that takes days in a spreadsheet is turned into an interactive Q&A session.
Upload paper PDFs → Dify Knowledge base (split by paragraph, embed) → Chatbot or Agent app → System prompt: "Answer from the provided literature only; cite paper title and page."
Claude Code ScientistPoint Claude Code at a 1 500-line Python analysis module with no tests. It reads the module, infers the expected behaviour of each function, and writes a pytest suite with edge-case coverage — all in the existing repo, no new scaffolding.
Coverage goes from 0 % to 87 % in one session, and the tests immediately surface two silent numerical bugs in the normalisation step.
Run `claude` in the repo root, then: "Write pytest tests for src/analysis/plate_normalise.py. Match the existing pytest conventions in tests/. Run the suite and fix any failures."
A 2 000-line data-cleaning script is split by Codex into four focused modules (ingest, validate, transform, export), with a clean public API and no change to observable behaviour — verified by re-running the original outputs.
The script becomes maintainable and reusable; a colleague adapts the ingest module for a new instrument within an hour.
Run `codex` in your project: "Refactor scripts/clean_sequencing_data.py into a package with ingest, validate, transform, and export modules. Preserve all existing behaviour. Add docstrings."
Opencode walks every `.py` file in the package, infers types from usage, adds PEP 484 annotations, and runs mypy in strict mode — iterating until the package passes with no errors and no `Any` suppressions in core logic.
A 6 000-line bioinformatics package gains full type coverage; mypy catches a silent integer-vs-float bug on the first run.
Run `opencode` in the package root: "Add strict PEP 484 type hints to every function and class in src/. Run mypy --strict and fix all errors. Do not add # type: ignore comments."
Given five real example CSVs with irregular headers, merged cells exported as blanks, and inconsistent date formats, Opencode writes a parser that handles all variants, returns a tidy DataFrame, and raises descriptive errors on unrecognised formats.
Manual pre-processing of instrument exports is eliminated; the parser handles every file variant from the past three years without modification.
Run `opencode`: "Write a Python parser in src/io/instrument_parser.py for the CSV format in data/examples/. Handle the header variants shown across the five sample files. Return a tidy pandas DataFrame."
A 1 800-line R script that does everything from data ingestion to figure output is split into sourced modules (ingest.R, clean.R, model.R, figures.R) with a thin orchestration script — keeping every output bit-identical to the original.
Each stage can be developed and re-run independently; the figures module is reused in a second paper within the same week.
Open the script in Antigravity: "Split this monolithic R script into ingest.R, clean.R, model.R, and figures.R modules. Add a run_all.R orchestrator that sources them in order. Verify all figure outputs are unchanged."
Copilot ScientistCopilot autocompletes pandas groupby chains, suggests the correct scipy statistical test given the data shape, and fills in markdown annotation cells explaining each step — all inline in VS Code as the notebook is written.
A full exploratory analysis that would take half a day is drafted in 90 minutes, with method documentation already in place for a manuscript methods section.
Open the .ipynb in VS Code with Copilot enabled. Start typing `df.groupby(` — accept completions, then add a markdown cell starting with `## Results` and let Copilot suggest the interpretation text.
Copilot Scientist CreatorIn VS Code, Copilot Chat explains the existing (broken) regex, proposes a corrected pattern for the messy log format, generates a battery of test strings covering edge cases, and adds an inline comment explaining the pattern in plain English.
A regex bug that stalled the ingestion pipeline for two days is identified and fixed in 30 minutes with a clear explanation the whole team can follow.
Select the broken regex in VS Code, open Copilot Chat (Ctrl+Shift+I): "This regex is not matching dates in the format DD/MM/YYYY HH:MM:SS. Fix it and explain each group. Add 10 test strings covering edge cases."
OpenClaw ScientistOpenClaw polls a public data API or FTP endpoint every few hours, compares checksums against the last known version, and fires a Telegram message the moment a new dataset release is detected.
You are the first to pull fresh data for your pipeline — no more daily manual checks or missed release windows.
In Telegram, send: "Watch [endpoint URL] every 4 hours. When the file changes, download it to ~/data/latest/, run my preprocessing script, and ping me with the diff summary."
OpenClaw Scientist AI-native SMEEach morning OpenClaw searches preprint servers and curated RSS feeds for your saved keyword list, ranks results by relevance using your local model, and sends a digest to Telegram with one-line summaries and links.
A focused reading list lands before you open your laptop — no manual search, no information overload.
Send to your OpenClaw bot on Telegram: "Every weekday at 7:30am, search arXiv and bioRxiv for my keyword list, pick the 5 most relevant, and message me a digest. Learn which ones I open."
Perplexity ScientistAsk Perplexity a focused research question and it answers with inline citations you can click through to the original papers — a fast, sourced orientation before a deep dive.
A starting map of the literature with real links, not a black-box summary.
Ask: "What are the leading hypotheses for X, with citations from the last 3 years?" then open each cited source.
ChatGPT ScientistPaste a broken Python/R snippet into ChatGPT, describe the error, and get a corrected version with an explanation of what went wrong.
A working script and an explanation you can learn from — no install needed.
Paste your code + the error message → "Fix this and explain the bug in one paragraph."
LM Studio ScientistA researcher loads a folder of de-identified clinical notes into LM Studio's local OpenAI-compatible server, then queries it in a simple chat UI to extract specific findings — adverse events, dosing records, outcome dates — without any data touching a cloud endpoint.
Ethics-compliant analysis of identifiable data is possible on a standard laptop, even before IRB-approved anonymisation is complete.
In LM Studio, download a medical-tuned model (e.g. Meditron or a Mistral variant), start the local server, then chat: "From the notes below, extract all adverse events and their dates." Paste notes into the message.
A field researcher with no reliable internet installs Jan on a laptop before departure, downloads a capable model, and uses it throughout fieldwork to interpret sensor readings, draft field notes in structured format, and answer methodology questions — entirely offline.
Research work continues productively in remote locations with no connectivity, and all raw observations stay on the device.
Open Jan, download a capable model (e.g. Llama 3 8B or Gemma 2 9B) from the model hub before you leave. In the chat, paste raw field observations and prompt: "Organise these into a structured field note with date, location, observation, and next steps."
A researcher drops a folder of papers — including unpublished manuscripts and licensed PDFs — into an AnythingLLM workspace and points it at a local model (via Ollama or LM Studio). They then ask cross-paper questions and get answers with citations back to the exact source document, with nothing ever leaving the laptop.
A grounded, citeable assistant over sensitive or licensed papers, with zero risk of the PDFs reaching a cloud provider or training set.
Install AnythingLLM, set the LLM provider to a local model (Ollama / LM Studio), create a workspace, drag in your PDFs, then ask: "Across these papers, what methods were used to measure X, and which reported effect sizes? Cite each source."
Claude Scientist AI-native SMEBefore pasting any unpublished results, patient measurements, or proprietary formulas into Claude, navigate to Claude.ai → Settings → Privacy → and confirm that "Improve Anthropic's products" is off (free tier) or use the API with no retention by default.
Your unpublished data is not used to train future model versions, protecting both research priority and participant confidentiality.
Check Claude's current model card or privacy settings page for the exact toggle name — it changes between product versions. The API has no human-review data retention by default.
ChatGPT Scientist AI-native SMEWhen ChatGPT provides a literature citation, paste the exact title and author list into Google Scholar, PubMed, or Semantic Scholar before including it in any document. Ask ChatGPT to provide the DOI and then resolve the DOI yourself.
Fabricated references — a well-documented failure mode of all large language models — are caught before they reach a report, grant application, or manuscript.
Follow-up prompt: "Give me the DOI for each paper you cited." Then open doi.org/<DOI> directly. If the DOI does not resolve, the citation is fabricated.
ChatGPT ScientistOpting out of model training is not the same as zero data retention — OpenAI's API and consumer tiers still transmit data to their servers and retain it for abuse-monitoring purposes. Anonymise or aggregate data before asking for analysis; never paste names, dates of birth, or medical record numbers.
You stay compliant with data-protection regulations (GDPR, HIPAA-equivalent obligations) that apply regardless of the provider's training opt-out.
Before sharing any dataset, run a quick de-identification step: replace names with "Participant A/B/C", dates with relative offsets (Day 0, Day 7), and IDs with sequential integers.
Gemini Scientist FounderEnable the "Google Search" grounding tool in Gemini Advanced so responses cite live web pages. For every cited link, open it and confirm the supporting sentence appears verbatim or paraphrase — do not assume the model quoted accurately.
Grounded responses are more likely to be traceable, but models still misquote or cherry-pick context; manual spot-checking catches these errors before they propagate.
In the Gemini web app, click the globe icon to enable Search grounding. After the response, click each citation chip and find the exact supporting sentence in the source page.
A Flowise chatflow loads a folder of PDFs into a vector store, then answers questions grounded in them — "Which of these papers used a knockout mouse model?" — with citations back to the source document.
A private research assistant that searches your own literature instead of the whole internet.
New Chatflow → PDF/Folder loader → embeddings → vector store → Conversational Retrieval QA node → ask a question in the chat panel.
A Langflow agent is given a "fetch from a public API" tool and a calculator tool, so it can answer "Pull the latest sequence record for accession X and tell me its length" by choosing the right tool live.
A reusable bench assistant the whole group can query — exported as an API endpoint.
New Flow → Simple Agent template → add an API Request tool component → open the Playground and ask a data question.
A CrewAI crew of three agents — a "searcher", a "summarizer", and a "critic" — divides a literature review: one finds papers, one extracts findings, one flags weak evidence, producing a structured draft.
A first-pass review draft assembled by agents collaborating, ready for you to verify and refine.
pip install crewai → crewai create crew lit_review → define researcher/writer/critic agents in the YAML → crewai run.
Base44 Scientist Creator AI-native SMEA tool to build a survey, collect responses from a public link, and analyze results on a live charts dashboard.
Own your survey data end to end with custom analysis instead of a locked-in form tool.
Build a survey app where I create questions (multiple choice, rating, text) and share a public link. Collect responses into a database and show a live dashboard with charts per question and a filterable table of raw responses I can export.
Lovable ScientistA portal where authors submit abstracts and organizers manage review and acceptance.
Run a small conference’s submissions without clunky academic software.
Build a conference abstract portal where authors log in and submit an abstract with title, co-authors, topic, and a file upload. Let organizers see all submissions, assign a status of Under Review, Accepted, or Rejected, and email authors the decision.
v0 Scientist AI-native SMEA data-viz dashboard summarising survey responses with charts, distributions and a respondents breakdown.
A clear visual readout of your study or feedback data.
Build a survey results dashboard with summary stat cards, a Likert-scale stacked bar chart, a demographics donut chart, a response-over-time line chart, and a filterable table of individual responses.
Ingests a raw Qualtrics export, recodes reversed items, filters incomplete responses, calculates composite scores, and outputs a publication-ready summary table with descriptive statistics.
Data cleaning that usually takes a day in SPSS or R is a reusable, auditable KNIME workflow.
CSV Reader → Row Filter (completion threshold) → Math Formula (recode reversed items) → GroupBy (mean, SD per scale) → Normalizer → Data to Report node.
Cursor reads the existing survey data schema and a static results page, then adds Recharts bar and Likert-scale charts that pull live data from the API — matching the existing Tailwind design system and responsive layout.
Survey results are visual and interactive instead of a wall of numbers; stakeholders explore the data directly without needing an export.
In Cursor chat: "Add Recharts bar charts and a Likert-scale heatmap to pages/results.tsx. Pull data from the existing /api/survey-results endpoint. Match the Tailwind classes already used on the page."
Lovable ScientistA team library where members save papers with tags, notes, and a shared reading list.
Keep the group’s literature in one searchable place instead of scattered PDFs.
Build a shared citation library where lab members log in and add papers with title, authors, year, link, tags, and notes. Make everything searchable and filterable by tag, let people mark papers as read, and show a feed of recently added papers.
v0 ScientistA searchable, filterable interface for a list of papers, with tags by topic, year and author.
A browsable paper list that beats a static reference dump.
Build a publications explorer UI: a search bar, filters for year, topic and author, and a list of paper cards each showing title, authors, venue, year, and links to PDF and DOI. Sortable by date or citations.
Hermes Scientist CreatorA personal knowledge base on a small VPS that the agent — not you — maintains, with a chat-bot interface and a static site.
Your notes compound over time instead of rotting, because something actively tends them.
Set up a self-maintaining research wiki on my VPS — watch my notes folder, update wiki pages when I add or edit files, and expose a chat interface so I can query it on Telegram.
Hermes Scientist CreatorStructured markdown notes in a synced Obsidian vault act as durable memory that survives context resets (a widely-upvoted community pattern).
The agent’s knowledge lives in files you own and can read, not a black box.
Use my Obsidian vault at [path] as your long-term memory — read existing notes at session start, write a structured summary note after each conversation, and sync it so my vault stays current.
Hermes ScientistA knowledge substrate that builds a qualitative "map" over existing Obsidian/vimwiki notes and sessions.
Your scattered notes become a navigable structure instead of a flat pile.
Read my Obsidian notes at [path], build a semantic map of the concepts and how they connect, save it as a skill, and let me query it by asking things like 'what links to [topic]?'
n8n Scientist AI-native SMEWatches an arXiv RSS feed for a query category, extracts title, authors, abstract, and PDF link, and appends each new preprint as a structured row in a shared Notion database.
The team's Notion library is updated daily with relevant preprints so no one misses a key paper.
RSS Feed trigger (arXiv category feed) → Set node (map title, authors, abstract, URL) → Notion node (create page in database).
Dify ScientistA RAG chatbot built on the lab's own protocol PDFs and SOPs. Researchers ask questions in natural language and get precise, cited answers drawn from the internal document corpus.
New lab members find protocol answers in seconds instead of hunting through shared drives or asking a senior colleague.
Upload PDFs (SOPs, protocols) → Dify Knowledge base (chunk + embed) → Create Chatbot app → Configure retrieval (top-k, similarity threshold) → Embed on internal wiki.
Windsurf reads every module in a bioinformatics pipeline, writes NumPy-style docstrings for all public functions and classes, configures Sphinx with autodoc, and verifies the docs build without warnings.
A 4 000-line pipeline that was a black box to new lab members now has searchable, navigable HTML documentation.
In Windsurf: "Write NumPy-style docstrings for every public function in src/pipeline/. Set up Sphinx with autodoc in docs/. Run sphinx-build and fix any warnings before finishing."
Claude ScientistUpload a 30-page PDF to Claude and ask it to extract the methods, list every assumption, and flag where the conclusions outrun the data.
A structured critique of a long document in minutes, holding the whole paper in context.
Drag in a PDF → "Summarize the method, list assumptions, and note where claims exceed the evidence."
A writer or researcher pastes their accumulated notes — drafts, excerpts, references — into Jan's chat and asks it to identify recurring themes, suggest a structure, or extract quotable sentences, with everything processed entirely on their own machine.
A private thinking partner that handles years of personal notes without any of that content going to an external server or training dataset.
Open Jan, select a mid-size model, paste your notes and ask: "What are the three strongest recurring themes here? List supporting quotes from the text for each."
Claude ScientistAsk Claude to draft a one-paragraph methods disclosure stating which tasks used AI assistance, which model was used, and that all outputs were verified by the authors — ready to paste into a manuscript.
Readers and reviewers know exactly where AI was involved, satisfying most journal integrity policies without guessing at the right wording.
Prompt: "Draft a transparent AI-use statement for a methods section. I used Claude to help structure the literature review and draft the discussion. I verified all claims against primary sources. Include the model name and that no data were shared."
Claude Scientist CreatorAfter Claude produces a draft section, send a follow-up: "List every factual claim in the above draft that you are less than fully confident about, and flag where a reader should verify against a primary source."
You catch hallucinated statistics and unsupported assertions before they enter your manuscript, instead of relying on Claude's confident tone as a proxy for accuracy.
Treat the critique list as a mandatory checklist — do not finalise the section until each flagged claim is verified in a primary source you can cite.
ChatGPT Scientist Creator AI-native SMEOpen every research-focused session with a system-level instruction: "If you are uncertain about any fact, say 'I'm not sure' and tell me what you are basing the answer on. Never produce a confident-sounding answer when your confidence is low."
The model's confident tone — its most dangerous trait for factual work — is partially counteracted, and you are more likely to receive explicit uncertainty signals you can act on.
Paste this instruction as the first message in every new conversation used for literature or factual research tasks.
Base44 Scientist AI-native SMEA tracker for grant applications with funder, amount, submission deadline, status, and assigned team member.
Never miss a funding deadline again — the team sees every application’s stage at a glance.
Build a grant tracker where I add funding opportunities with the funder name, amount, deadline, and status (Drafting, Submitted, Awarded, Rejected). Assign each grant to a team member, give everyone a login, and show a calendar of upcoming deadlines.
Lovable ScientistA tool to track grant applications, deadlines, reporting milestones, and assigned owners.
Never miss a submission or progress-report deadline again.
Build a grant tracker where lab members log in and add grant applications with funder, amount, deadline, status, and the person responsible. Show an upcoming-deadlines view, send reminders as deadlines approach, and let me filter by status and owner.
v0 ScientistAn internal tool to track grant applications with status, deadlines, amounts and a summary overview.
A tidy way for a lab to see funding pipeline at a glance.
Build a grant tracking tool with summary cards (total awarded, pending, deadlines this month), a table of applications showing funder, amount, status badge and deadline, and filters by status and year.
Hermes ScientistPulls Whoop fitness data locally and exposes Hermes tools so you can ask custom questions over your own biometrics.
Your health data stays on your machine while still being queryable like a dataset.
Connect to my Whoop data locally, build tools for HRV, strain, and recovery trends, then schedule a morning message on Telegram each day with yesterday’s key metrics and any anomalies.