Tools Reference
When you connect your AI agent to Lomita, you see 9 tools. These are all you need — your research team handles everything else behind the scenes.
Your Tools
| Tool | What it does |
|---|---|
research | Ask a research question and get a data-backed answer |
follow_up | Dig deeper into a previous finding |
status | Check on your active research |
sources | Search 104+ data sources |
upload | Add your own data (CSV, JSON, Parquet) |
monitor | Watch a research question over time with scheduled re-analysis and alerts |
deliver | Send reports to email, Slack, Discord, Teams, Zulip, or any webhook |
set_provider | Configure your LLM provider |
list_providers | See which providers are configured |
pause_agents | Pause all research agents to save LLM tokens |
resume_agents | Resume paused agents |
How to use them
research
The primary tool. Ask any research question and the research team does the rest.
"Research whether consumer sentiment predicts retail sales growth"
What happens behind the scenes:
- Formalizes your research question
- Finds relevant data from 104+ sources
- Prepares and connects the data
- Runs statistical tests (correlation, lag analysis, regression, regime analysis)
- Compiles an executive report
- Report is published to your research repository
You get back a Research ID. Use status to track progress.
follow_up
Branch from an existing investigation to explore a related question.
"Follow up on research abc-123: break this down by pre-2020 vs post-2020"
Creates a follow-up investigation linked to the original, then runs the same research process.
status
Check progress on your research. Call with no arguments to see everything, or with a Research ID for details.
"Check the status of my research"
"What's the status of research abc-123?"
Returns: research status, connected data sources, recent activity, and a link to the report when complete.
sources
Search the data catalog to see what's available before starting research.
"What data sources do you have for energy markets?"
"Search for sentiment data"
Returns matching sources with descriptions, domains, and whether authentication is required.
upload
Add your own proprietary data to the data lake. Returns a presigned upload URL.
"Upload my-sales-data.csv"
Your data is stored privately and becomes available for analysis alongside the 104 public sources.
monitor
Set up continuous monitoring on a research question so it is automatically re-analyzed on a schedule.
"Monitor research abc-123 weekly and send results to my email and Slack"
Parameters:
| Parameter | Description |
|---|---|
hypothesis_id | Research ID — the question to monitor |
frequency | How often to re-analyze: daily, weekly, or monthly |
channels | Where to deliver results: email, webhook, or both |
enabled | Set to false to disable monitoring |
How monitoring works:
- On each scheduled cycle, the research team re-runs the full analysis against the latest data.
- Reports are always delivered — even when findings are stable. Absence of change is valuable information.
- If a research status flips (e.g., SUPPORTED to INCONCLUSIVE), an immediate alert is sent outside the normal schedule.
- Updated reports are committed to your research repository as new versions. Use
git diffto see exactly what changed between cycles.
To disable monitoring:
"Stop monitoring research abc-123"
See Continuous Monitoring for a full walkthrough.
deliver
Send a completed report to email, a webhook, or both. Auto-detects the platform from the webhook URL.
"Deliver report abc-123 to [email protected]"
"Deliver report abc-123 to https://hooks.slack.com/services/T00/B00/xxx"
"Deliver report abc-123 to [email protected] and https://discord.com/api/webhooks/123/abc"
Supported destinations:
| Destination | How to specify |
|---|---|
| Provide an email address | |
| Slack | Provide a Slack incoming webhook URL |
| Discord | Provide a Discord webhook URL |
| Microsoft Teams | Provide a Teams incoming webhook URL |
| Zulip | Provide a Zulip webhook URL |
| Generic webhook | Provide any HTTPS URL |
The platform is auto-detected from the URL — no extra configuration needed. You can provide both an email and a webhook URL in a single call.
set_provider
Configure the LLM that powers your research agents. This is typically the first thing you do after connecting.
"Set provider to zen with key sk-abc123..."
"Configure anthropic with my API key"
Supported providers: Anthropic, OpenAI, Google (Gemini), DeepSeek, Zen (OpenCode), Groq, Mistral, xAI, OpenRouter, Cerebras.
Zen tiers:
| Tier | What you get |
|---|---|
zen | Standard pay-per-token via OpenCode |
zen-go | Unlimited subscription — uses Kimi K2.6 and DeepSeek V4 Pro with no per-token charges |
To use the unlimited tier:
"Set provider to zen-go with my subscription key"
When you set a provider, your research team is automatically configured and ready to work.
list_providers
See which LLM providers you've configured.
"List my configured providers"
pause_agents
Pause all research agents to stop LLM token consumption. Useful when you're done for the day or want to review results before continuing.
"Pause all agents"
Any in-progress work is suspended gracefully. Agents can be resumed at any time.
resume_agents
Resume previously paused agents to continue research.
"Resume agents"
Agents pick up where they left off.
What you don't see
Behind your 9 tools, there are 44+ internal tools that the research team uses. These handle data connections, SQL queries, code review, task management, and coordination. You never need to touch them — the research tool orchestrates everything automatically.
Reports
When research completes, the report is published to your research repository:
https://YOUR-INSTANCE-git.lomita.io/lomita/research
Each investigation gets its own directory with:
report.md— the executive report (rendered as formatted markdown)analysis/— SQL queries and Python scripts useddata-sources.md— what data was used
The repository's README shows an auto-generated index of all your research with status and links.