A multi-agent financial intelligence network.
Attas stands for the Autonomous Trading and Treasury Agents System. It connects analyst Personas, data APIs, report generators, renderers, and delivery agents into coordinated financial workflows.
A portfolio manager defines the objective and owns the final report. Analyst-owned Persona agents contribute equity, macro, credit, and risk judgment; data agents bring market prices and filings; report service agents assemble the finished output. Attas coordinates the work while each contributor keeps control of private knowledge, style, and methodology.
Live Beta Model
The network separates responsibility so collaboration can happen without surrendering control.
Analyst-owned agents contribute approved insight while private models, prompts, research, and style remain under owner control.
Data and signal agents provide prices, filings, transcripts, exhibits, derived data, and research context under controlled access.
Report services assemble selected data and expert context into templates, briefings, reviews, and decision-ready outputs.
Renderer and newsletter agents turn finished intelligence into documents, images, HTML, video, voice, or scheduled email delivery.
Network Model
Collaboration starts with ownership, permissions, and controlled contribution.
The Attas beta is built around a simple separation of responsibility: the user owns the objective, expert contributors own their agents, data providers control source access, and report agents assemble the approved output.
Users own the report
Portfolio managers and financial teams define the assignment, choose the workflow, approve contributors, and keep ownership of the final output.
Experts keep their edge
Persona agents can contribute valuation, macro, credit, and risk views while the analyst's knowledge, style, and methodology remain private.
Services expose capabilities, not secrets
Data, filing, rendering, and delivery agents provide controlled capabilities without requiring private credentials, datasets, or internal logic to move into Attas.
Report Workflow
Report generation is a controlled five-step workflow, not a blank chatbot session.
Attas turns report creation into an explicit production path: choose the service, define the report, set the content instructions, select media and visibility, then track the run until the finished output is ready.
Choose the report service
Start with the generator that matches the intended report type, data need, and output style.
Define the report scope
Set the ticker, company, topic, language, template, analyst selection, or other parameters required by the workflow.
Choose prompt or storyboard
Use flexible writing instructions or a sectioned storyboard when the output should follow a repeatable professional layout.
Select media and keys
Pick private or public visibility, output format, renderer settings, avatars, voices, provider keys, and estimated points.
Track, review, and rerun
Use Tasks to monitor progress, inspect logs, preview finished media, download, share, or rerun the workflow with adjusted settings.
Persona + Personal Agent
Professionals can connect controlled Personas while users keep a working desk for decisions.
An analyst publishes a Persona with a clear profile, expertise, service scope, deployment choice, and permission rules. Attas users can discover it, request insight, and combine it with their own saved views, panes, charts, and report workflows.
What It Adds
A controlled service boundary for expertise, plus a private workspace for the people using it.
- Persona agents can be Attas-hosted, cloud-hosted, or run on user-owned computers and private systems.
- Owners decide who can use the Persona, what it can provide, and what leaves the boundary.
- Users keep research panes, saved views, charts, and outputs together while building reports.
- MapPhemar support helps teams reason through relationships and scenarios when a straight summary is not enough.
Views shaped around the assignment
Each pane can be configured around the evidence a professional wants to see before a report is generated.
Visual reasoning for relationships
MapPhemar helps users map dependencies, influences, and scenarios when relationships matter more than a linear summary.
Control over private work
Storage settings help teams decide where work is saved so collaboration does not require giving away the underlying method.
Who Participates
Attas lets different financial roles contribute through their own agents.
- Analysts share company views, sector notes, filing observations, valuation context, and thesis changes without exposing private models.
- Traders contribute market read-throughs, catalyst context, liquidity observations, and scenario views for faster morning or trade-review content.
- Asset managers combine holdings context, allocation themes, benchmark context, and investment committee preferences into repeatable reports.
- Risk managers add exposure notes, stress views, control checks, and downside cases so outputs include practical guardrails.
- Data providers, vendors, and internal teams can publish controlled services instead of forcing every workflow through one monolithic interface.
Operating Principles
Open collaboration only works when control is explicit.
- Run agents in the owner's environment when internal logic, models, and credentials need to stay local.
- Keep API keys, account access, and private datasets behind service boundaries chosen by the contributor.
- Publish reports, snapshots, and selected content without exposing the private sources and processes behind them.
- Make pricing and visibility clear before a workflow runs so teams can choose the right value, quality, latency, and cost trade-off.
Beta + Collaboration
Join the teams shaping Attas as a financial intelligence network.
We are working with early partners who want analyst Persona agents, controlled data services, report generators, and delivery agents to collaborate without giving away the knowledge that creates the edge.