ARS-native shell design
Design tables, figures, and listings as Analysis Display Metadata aligned to the CDISC Analysis Results Standard.
From ARS-driven TFL shells to production-ready SAS and R code in one connected workspace.
Generate production-ready SAS and R code from ARS-driven TFL shells. TFL Designer + Code Generators turns approved Tables, Figures, and Listings shells into executable programs with matching QC scripts.
Approved SAP metadata defines the analyses that the shell set must cover.
Shell development, review, updates, and approval drop from ~120 hours to ~18 hours per study.
First-draft TFL output development and QC falls from ~16 hours per output to ~6.5 hours per output.
Production-ready SAS and R code is generated directly from approved, ARS-aligned shells.
Shells become structured metadata, and that metadata drives both the mock-ups your team reviews and the code your programmers run.
Design tables, figures, and listings as Analysis Display Metadata aligned to the CDISC Analysis Results Standard.
Auto-populate realistic mock-up shells so reviewers see exactly what the final output will look like.
Generate first-draft, production-ready SAS and R programs directly from approved, ARS-aligned shells.
Produce companion QC programs alongside every generated production program for independent verification.
Reuse a growing library of shell templates, including FDA Safety Tables and Figures Integrated Guide (STF-IG) templates.
Produce the TFL table of contents and export shells to RTF, PDF, XML, and JSON for any downstream consumer.
Track shell changes, review with stakeholders, and approve with e-signature before programming begins.
TFL Designer + Code Generators sits between the approved SAP and the review workspace, turning signed-off shells into executable SAS and R programs.
Analyses approved in SAP Builder generate the initial shell set with populations and parameters already attached.
Lay out each table, figure, and listing as structured display metadata with titles, footnotes, and formats.
Circulate mock-ups, capture comments, version the shells, and sign off before any code is written.
The AI Code Generator produces first-draft SAS and R programs plus matching QC scripts against the approved shells and ADaM data.
Run the generated code inside the SCE, review outputs in TFL Viewer, and regenerate automatically when shells or data change.
Explore a free, web-based version of TFL Designer with a growing library of community templates, FDA STF-IG shells, and CDISC ARS standards.
A secure, cloud-based module of Clymbr Hub , governed, versioned, and audit-ready from day one.
Runs inside Clymb's validated Statistical Computing Environment with role-based access, encryption in transit and at rest, and elastic compute for large output sets.
Select any capability to see what it does.
Production-ready SAS and R code generators from ARS-driven TFL shells are automation tools that read CDISC Analysis Results Standard (ARS) metadata for Tables, Figures, and Listings and write executable SAS or R programs directly from that metadata. Instead of manually translating a static shell into code, statistical programmers start with first-draft programs that already reflect the approved display structure, analysis populations, parameters, and sponsor macro libraries. Clymb Clinical's TFL Designer + Code Generators module produces these programs inside a validated Statistical Computing Environment, with matching QC scripts generated alongside every production program.
TFL shells are stored as structured Analysis Display Metadata, not documents, so every title, footnote, and analysis definition is machine-readable.
The same approved shell can generate production-ready SAS programs, R programs, or both, keeping mixed-language teams on one specification.
Each generated production program is paired with a matching QC script for independent double-programming verification.
Because the shells, code, and QC scripts all share the same ARS metadata, specifications and programs never drift apart. Changes to the shell propagate automatically to the generated code, cutting TFL output programming and QC time by 60% or more.
Three ways to turn approved TFL Designer shells into executable code. All three read the same CDISC ARS metadata, so you can mix them across a programme without re-specifying a single output.
Agentic AI · inside TFL Designer
AI Code Generator lets users generate their SAS or R code directly within TFL Designer in a matter of seconds, eliminating manual programming and dramatically accelerating output creation and QC workflows.
| Attribute | AI Code Generator | siera | atlas |
|---|---|---|---|
| Output language | SAS or R | R | SAS |
| How it works | Agentic AI drafts code from the approved shell | Metadata-driven R scripts and ARDs | Metadata plus macro library into SAS programs |
| Standard | CDISC ARS + sponsor-specific standards | CDISC ARS + sponsor-specific standards | CDISC ARS / ARD + sponsor-specific standards |
| Best for | Probabilistic approach , speed across mixed SAS and R teams | Deterministic approach , open-source R teams and ARD generation | Deterministic approach , established SAS macro environments |
| Availability | Clymbr Hub module | Open source (GitHub, Pharmaverse) | Clymbr Hub Module or SAS |
| Maturity level | Mature | Very Mature | Very Mature |
They are automation tools that read CDISC Analysis Results Standard (ARS) metadata for Tables, Figures, and Listings and write executable SAS or R programs directly from that metadata. The generated code reflects the approved display structure, analysis populations, parameters, and sponsor macro libraries, so programmers start from a working first draft instead of a blank editor.
TFL Designer stores shells as structured Analysis Display Metadata aligned to CDISC ARS. The Code Generators read this metadata, along with ADaM dataset structures and sponsor macro libraries, and emit first-draft SAS or R programs that implement the defined titles, footnotes, populations, and statistical methods.
The code is generated against approved, ARS-aligned shells and sponsor-standard macro libraries, runs inside your validated Statistical Computing Environment, and is paired with auto-generated QC scripts. It follows CDISC, SAS, and R best practices and can be reviewed, amended, and locked under version control like any other production program.
Yes. The same approved shell can produce SAS programs, R programs, or both. This keeps mixed-language teams on one specification and eliminates the need to maintain separate shell definitions for each language.
The AI Code Generator uses agentic AI to draft SAS or R programs from approved shells inside TFL Designer. siera is an open-source, Pharmaverse-approved R package that turns ARS metadata into R scripts and Analysis Results Datasets. atlas is a SAS-based automation engine that converts ARS metadata and a validated macro library into executable SAS programs. All three read the same CDISC ARS metadata, so teams can mix them across a programme.
Yes. Every generated production program is accompanied by a matching QC script. This supports independent double-programming verification and keeps the QC process aligned with the same ARS metadata used to create the production code.
It is built on CDISC Analysis Results Standard (ARS) and Analysis Display Metadata (ADM), and interoperates with SDTM and ADaM. FDA Safety Tables and Figures Integrated Guide (STF-IG) templates are included in the community version, and generated code aligns with SAS and R best practices.
Generated SAS and R programs execute inside your Statistical Computing Environment (SCE), whether that is a validated cloud instance, on-premises setup, or sponsor-specific compute target. This keeps execution within your controlled, audit-ready infrastructure.
Bring your team, your study, and your code. We'll show you TFL Designer + Code Generators working end-to-end alongside the rest of the platform.